Tesi sul tema "Sensitive Questions"

Survey methodologists have given considerable attention to minimizing the response effects associated with asking sensitive questions. However, this work is largely atheoretical; little attention is paid to issues of definition, rather a common sense understanding of what is sensitive is assumed. Such research largely focuses on how sensitive behaviours are reported and provides little insight into the relationship between responses to attitude items and their sensitivity. Recent work of cognitive psychologists, although drawing attention to the way in which responses are affected by the immediate context of the questionnaire, does not extend to suggesting whether questions that vary in their sensitivity are likely to be differentially affected by this. In the light of these issues five studies are reported that suggest the importance of taking the sensitivity of attitude items into account. Sensitive items are defined in relation to the notion of 'threat' and the context dependence of what is considered to be sensitive is recognised. The sensitivities of a pool of attitude items for a particular population is thus established in advance of being used in further work. The different patterns of responses associated with high and low sensitive items are noted and are seen to bear an orderly relationship with agreement ratings that is not explicable in terms of the orientation of the items. Results also indicate that items that varied in their sensitivity are differentially affected by manipulations in the context of threat within which the questionnaire was embedded. The nature of these effects is explored as is the notion that responses to sensitive items may be used to cope with threat. The implications of these results for research using attitude items within questionnaires is discussed. It is suggested that knowing the sensitivities of the items being used is a necessary first step in being able to determine the extent to which these affect the results.

This dissertation explores methods to improve the quality of data about sensitive labor market topics, such as undeclared work and receipt of basic income support in Germany, using surveys of the general population. Due to the sensitive nature of both topics, respondents may choose to misreport and adjust their answers in accordance with social norms. Over the past decades, special strategies—particularly targeted to reduce misreporting on sensitive topics—have been developed. One such class of data collection strategies are so-called ‘dejeopardizing’ techniques, out of which the randomized response technique (RRT) and the item count technique (ICT) are the most popular and best investigated ones. The goal is to elicit more honest answers from respondents by increasing the anonymity of the question-and-answer process. These techniques provide prevalence estimates as well as estimates of regression coefficients, regressing dependent variables generated by means of RRT or ICT on a set of covariates of interest. However, these dejeopardizing techniques have not been applied to collect data on undeclared work or receipt of welfare benefits in German surveys. This dissertation aims at closing this gap using an experimental design that allows us to compare the performance of these dejeopardizing techniques to direct questioning. In 2010 we conducted two telephone surveys on undeclared work and welfare benefit receipt. We experimentally tested whether the RRT, the ICT, or the newly developed item sum technique (IST) reduce bias due to social desirability compared to direct questioning (under the ‘more-is-better’ assumption and using validation data in one study). Our results suggest that neither the RRT nor the ICT provide unambiguous results with respect to improving the accuracy of reports of the socially undesirable behavior, while the IST results were more promising. This dissertation provides insights into a variety of practical and theoretical factors contributing to a successful implementation of the RRT, the ICT and the IST in labor market surveys.
Die vorliegende Dissertation geht der Frage nach, wie das Ausmaß von Schwarzarbeit und Arbeitslosengeld-II-Bezug in Deutschland im Rahmen von Befragungen der allgemeinen Bevölkerung möglichst valide geschätzt werden kann. Aufgrund des heiklen Charakters beider Themen ist davon auszugehen, dass Selbstauskünfte häufig nicht der Wahrheit entsprechen und stattdessen in vielen Fällen sozial erwünschte Antworten gegeben werden und das Verhalten systematisch unterberichtet wird. Um diesen Antwortverzerrungen entgegen zu wirken, wurden in den letzten Jahrzehnten in der empirischen Sozialforschung alternative Befragungstechniken entwickelt. So basieren beispielsweise die Randomized Response Technique (RRT) und die Item Count Technik (ICT) auf dem Prinzip der verschlüsselten Antworten und sollen durch eine Erhöhung der Anonymität in der Interviewsituation sozial erwünschtes Antwortverhalten reduzieren. Der Vorteil dieser Erhebungsverfahren liegt darin, dass zum einen weniger Annahmen hinsichtlich der Schätzungen getroffen werden und zum anderen mittels statistischer Auswertungen ziel gerichtet multivariate Zusammenhänge zwischen einer mit ICT oder RRT generierten abhängigen Variablen und Kovariaten auf individueller Ebene untersucht werden können. Bislang wurden diese Techniken allerdings noch nicht zur Erhebung von Schwarzarbeit oder des Bezugs von Arbeitslosengeld-II in Deutschland eingesetzt. Die Dissertation schließt diese Lücke und beschäftigt sich mit einem experimentellen Vergleich – sowie einer Weiterentwicklung – von Erhebungstechniken speziell für heikle Fragen mit einer direkten Befragung im Kontext von Arbeitsmarktsurveys. Mittels Fragen zum Thema Schwarzarbeit und zum Arbeitslosengeld-IIBezug, wird im Rahmen zweier Bevölkerungsbefragungen aus dem Jahre 2010 empirisch untersucht ob die RRT, die ICT bzw. die eigens entwickelte Item Sum Technik (IST) den Befragten tatsächlich ein höheres Ausmaß sozial unerwünschter Antworten entlocken als die direkte Befragung (unter der bekannten ‘more-is-better’ Annahme sowie mittels einer Validierungsstudie). Die Befunde zeigen, dass die häufig angenommene Wirkung der RRT oder der ICT auf die Bereitschaft der Befragten, sozial unerwünschtes Verhalten zu berichten, nicht eindeutig ausfällt. Die Ergebnisse der IST fallen hingegen positiver aus. Die vorliegende Dissertation liefert somit Hinweise hinsichtlich verschiedener praktischer als auch theoretischer Faktoren, die zu einer erfolgreichen Implementation der RRT, der ICT und der IST in Arbeitsmarktsurveys beitragen können.

Cette recherche montre l’écart entre les intentions politiques et institutionnelles et les pratiques des professeurs des écoles, qui sont partielles et hétérogènes. Il existe une pluralité de points de vue sur la légitimité et la pertinence de l’enseignement laïque des faits religieux, qui demeure un objet scolaire discuté. Un pan de notre étude est consacré à l’analyse des pratiques de professeurs enseignant les faits religieux pour éduquer à la laïcité. Ceux-ci, en prenant en charge l’enseignement de questions vives, ont dû faire preuve d’ingéniosité pédagogique, se traduisant par des bricolages rendus nécessaires par la rareté des outils, le manque de formation et l’absence de préconisations institutionnelles quant aux savoirs intermédiaires adéquats. Si, globalement, leurs pratiques répondent aux objectifs principaux de l’enseignement des faits religieux, notamment éduquer à la laïcité, aborder ces sujets suppose un investissement personnel lourd, pour l’enseignant comme pour ses élèves, et renvoie chacun à ses systèmes de valeurs. Les principaux écueils relevés sont liés au respect de l’obligation de neutralité (difficulté à se représenter sa pertinence pédagogique, sa transposition). En l’absence d’outils et de formation dédiés, certains enseignants sont démunis. L’association ENQUÊTE a été créée en 2010 pour répondre à ce besoin. Elle souhaite contribuer à réduire l’écart entre les intentions institutionnelles favorables à l’enseignement des faits religieux dans le cadre d’une éducation à la laïcité et les pratiques des professeurs, via des formations et des outils, et aussi à promouvoir cet enseignement et cette "éducation à" auprès des acteurs politiques et institutionnels
This research highlights a gap between political and institutional intentions, and practices of school teachers, which are partial and heterogeneous. There is a plurality of views about the legitimacy and relevance of teaching about religions and beliefs in a secular way, which is still a very much debated school subject. One part of this study is dedicated to the analysis of teachers' practices when teaching about religions and beliefs in order to educate to laicity. To address these sensitive education issues demands creativity. Because of a scarcity of tools, lack of training and very few institutional recommendations about adequate intermediate knowledge, teachers have had to cobble together their own teaching materials. Overall, their practices meet the main purposes of teaching about religions, especially to educate to laicity. Nonetheless, dealing with these subjects requires a heavy personal investment, for the teacher as well as for their students, and brings back each one to his own value systems. The main identified pitfalls are linked to the neutrality obligation : they find it difficult to represent their pedagogical relevance and to translate it into their teaching. Many teachers feel helpless when confronted to the lack of dedicated tools and training. ENQUÊTE is an association that was created in 2010 in order to address this issue. Its first objective is to contribute, via trainings and tools, to reducing the gap between institutional intentions that favour teaching about religions and beliefs – as part of the education to laicity – and teachers' practices. Its second goal is to promote this teaching among political and institutional actors

Генітивні речення розглядаються як односкладні субстанційні, у яких головний член представлений незалежною формою родового відмінка іменника, що може виражати наявність, існування, буттєвість предмета й водночас неозначено велику, надлишкову кількість або, навпаки, цілковиту відсутність речей. У їхній семантиці відсутні як значення процесуальності, так і динамічної ознаки та позачасової дії, а категоріальне значення буттєвості постає як граматичне значення констатації реальності в момент спілкування. Розкриваються структурно-семантичні моделі й комунікативно-модальні різновиди односкладних генітивних речень в мові друкованих ЗМІ.
Genitive sentences are considered to be one-member substantive ones, in which the main part is represented by an independent form of the genitive case of the noun, which can express the occurrence, existence of the subject, and at the same time undefined large, exuberant number or, conversely, total lack of things. In their semantics there is neither value of procedurality nor dynamic characteristic or timeless action, but the categorical significance of existentialism appears as the grammatical significance of the statement of reality at the moment of communication. Structural-semantic models and communicative-modal varieties of one-member genitive sentences in the language of print media are revealed.

碩士
銘傳大學
應用統計資訊學系碩士班
96
Social surveys are sometimes related to sensitive characteristics which respondents want to hide from others. It is difficult to collect reliable data from respondents and hard to improve the quality of surveys when the topic is sensitive. Warner (1965) proposed “Randomized Response Model” at first. After that, a lot of scholars proposed alternative randomized response models to improve Warner’s (1965) model. For example, Christofides (2003) proposed a new randomized response model. The respondent doesn’t answer 「yes」and「no」directly, and just reply a 「digit」. Huang (2004) proposed a new randomized response model which combines direct response model with Warner’s (1965) model. Kim & Huang (2004) proposed “A stratified Warner’s randomized response model” based on Warner’s model . In this article, we proposed an alternative model which combines Huang’s (2004) model with Christofides’ (2003) model and wishes supply enough protection to the respondents. And we can estimate the proportion of sensitive group more precisely with other models.

碩士
銘傳大學
應用統計資訊學系碩士班
95
In a sampling survey, if the topic of study is involved in sensitive characteristics, respondents usually refuse to reply or answer untruthfully. Warner (1965) proposed a “Randomized Response Model” to reduce the sense of insecurity of respondents and obtain an accurate population proportion estimator. Another procedure is for two unrelated sensitive questions proposed by Chang and Liang (1996). Their procedure can estimate two population proportions at the same time. Christofides (2003) proposed alternative randomized response Model. Respondents do not answer “yes” or “no” directly, reply “distance” instead in the procedure. Besides, Kim & Warde (2004) proposed stratified randomized response to enhance the efficiency of estimator. This paper combines Chang and Liang (1996) with Christofides (2003). Besides, Kim & Warde (2004) also is considered. Therefore we propose two new randomized response models and we also show the estimators of sensitive population proportion, the variances of estimator and simulating data for the variances of the proposed estimators will be presented. In addition, we can prove the variance of the proposed model is more efficient than that of Christofides (2003) Model.

Soziale Erwünschtheit ist ein in der sozialwissenschaftlichen Forschung bekanntes Phänomen. Es beschreibt die Tendenz von Befragten sich gegenüber dem Interviewer, Dritten oder einer fiktiven Öffentlichkeit in einem „guten Licht“ zu präsentieren, insbesondere wenn sensible Themenbereiche angesprochen werden. Trotz nunmehr rund 60 Jahren Forschung ist es jedoch bisher nicht hinreichend gelungen, das Ausmaß dieser Verzerrung methodisch zu erfassen. Das Forschungsfeld zu diesem Phänomen ist breit, aber genau hier liegt die Problematik. Viele der Studien beschäftigen sich mit isolierten Effekten der sozial erwünschten Antwortverzerrung (SD-Bias), ohne dabei weitere wichtige Einflussfaktoren zu berücksichtigen. Darüber hinaus werden in den Studien sehr unterschiedliche Dimensionen und/oder Operationalisierungen des SD-Bias angesprochen. Aus diesem Grund besteht der erste Abschnitt der vorliegenden Dissertation aus einer systematischen Auseinandersetzung mit den verwendeten Begrifflichkeiten, Operationalisierungen und theoretischen Ansätze im Rahmen der SD-Forschung. Da die meißten theoretischen Argumente auf dem Rational-Choice Ansatz basieren oder auf ihn zurückgeführt werden können, konzentriert sich die vorliegende Arbeit auf soziale Erwünschtheit im Kontext der Rational-Choice Forschung. Bei genauer Auseinandersetzung mit der Thematik wird deutlich, dass das Phänomen der „sozialen Erwünschtheit“ eine Reihe unterschiedlicher Dimensionen umfasst, die berücksichtigt werden müssen, um eine präzise Begriffsklärung zu ermöglichen. Je nachdem, welcher Aspekt betrachtet wird, verändert sich der Blickwinkel, aus dem heraus soziale Erwünschtheit analysiert wird. Um die Vergleichbarkeit der Forschungsergebnisse zu gewährleisten, ist es jedoch notwendig, das zu untersuchende Phänomen „soziale Erwünschtheit“ und seine jeweiligen Dimensionen so exakt wie möglich zu beschreiben. Unter welchen Bedingungen tritt nun sozial erwünschtes Antwortverhalten nach dem Racional-Choice Ansatz auf? Dem Ansatz von Hartmut Esser und seiner Weiterentwicklung durch Volker Stocké nach kann das Anerkennungsbedürfnis als Persönlichkeitskonstrukt nur zu einer Antwortverzerrung führen, wenn der Befragten eine Norm als sozial erwünscht wahrnimmt und sich nicht in einer völlig anonymen Situation befindet. Es konnte gezeigt werden, dass die verschiedenen Ansätze theoretische Schwächen haben, die bei der Umsetzung und Interpretation von Ergebnissen berücksichtigt werden müssen. Der unpräzise Umgang mit Begriffen und theoretische Lücken sind dabei besonders gravierend. Auch die Skalen, die zur Kontrolle des SD-Bias genutzt werden, zeigen unsystematische Bezüge zu den Dimensionen der Erwünschtheit. Gerade bei der Erwünschtheitswahrnehmung und den Täuschungsskalen wird deutlich, dass beim exakten Umgang mit der Operationalisierung verschiedene Ebenen des SD-Bias sichtbar werden, die bei der Analyse berücksichtigt werden müssen. Bisher ist nicht geklärt, ob die verschiedenen Messungen tatsächlich ein und dasselbe Konstrukt erfassen und im Rahmen der Erklärung des SD-Bias funktional äquivalent einsetzbar sind. Um das Forschungsproblem näher zu bestimmen werden empirischen Abschnitt die Ergebnisse von drei empirischen Studien präsentiert. Zunächst wird im ersten Schritt mittels Faktorenanalyse gezeigt, dass die sogenannten Lügenskalen nicht die theoretischen Erwartungen erfüllen. Da in Bevölkerungsumfragen trotz dieser Schwächen, die Skalen immer noch Verwendung findet, wird im darauffolgenden Schritt der Erklärungsbeitrag dieser Skalen auf Vorurteilsmessungen untersucht. Auch hier zeigen sich fast ausschließlich unsystematische Ergebnisse, die im Zusammenhang mit aktuellen Forschungsergebnissen den Nutzen der Täuschungsskalen als einzige Determinante zur Kontrolle des SD-Bias in Frage stellen. Im dritten Schritt wurde deshalb der Beitrag umfassenderer Modelle zur Erklärung von Vorurteilen untersucht. Die Daten basieren auf einer Studierendstichprobe. Die Befunde lassen erkennen, dass keines der Modelle die gewünschte Stabilität erreichte und zum Teil sehr empfindlich auf kleine Veränderungen in der Operationalisierung reagierten. Im abschließenden Kapitel wird festgehalten, dass die unsystematische Forschungslage in Bezug auf den SD-Bias eine Ursache unklarer theoretischer Konzepte und Operationalisierung ist. Des weiteren wird klar, dass eine große Anzahl zusätzlicher Messungen notwendig wäre, um den SD-Bias in einer Umfrage korrekt zu kontrollieren. Da bisher jedoch nicht geklärt ist, unter welchen Bedingungen der SD-Bias ein ernsthaftes Problem, gibt es keine Rechtfertigung für den kosten- und zeitintensive Aufwand dieser Messung im Rahmen von Befragungen. Um das Zusammenspiel der einzelnen Komponenten und die Folgen ihrer Messung korrekt bewerten zu können, wäre ein umfassendes Forschungsprogramm dringend erforderlich.

碩士
逢甲大學
統計學系統計與精算碩士班
107
In general, when questionnaires are used to collect the data on sensitive issues, most respondents tend to become hostile to the interviewers. The respondents are more likely to provide untruthful responses or refuse to respond. Subsequently, this makes the collect of data difficile and the analysis results meaningless. Fortunately, the use of a randomized response question can protect the privacy of the respondent and further motivate the respondent to provide truthful responses. In this regards, Xiao (2018) used Greenberg, et al. (1969) uncorrelated randomized response technique (RRT) to provide conditional posterior distribution under all potential variable information. He used the Gibbs sampling technique to provide the estimator of the proportion of sensitive attributes under different exogenous groups. Hsieh and Perri (2019) gave the Bayesian inference of the proportion of sensitive attributes by combining Gibbs sampling technique with the randomized response techniques proposed by Christofides (2003). In this paper, we extends the Bayesian inference of Xiao (2018) to the RRT proposed by Christofides (2003) and derive the Bayesian inference of the proportion of sensitive attributes under different exogenous groups. Moreover, we use the Bayesian method combined with Gibbs sampling to provide different models for the sensitive attributes under different exogenous groups. Simulation studies is conducted in order to evaluate the performance of the proposed method. In addition, we use the RRT of Christofides (2003) to collect the data of the Sexual EQ questionnaire conducted by the Student Affairs Office of Feng Chia University during the second semester of the academic year 2016. The respondents basic information and the online dating experiences were used as external variables to obtain the network one-night ratio estimation under different group combinations.

The research described in this thesis reviewed previous uses of Interactive Voice Response (IVR), developed appropriate software, and employed IVR to obtain self-report of sensitive issues in surveys and conduct brief public health interventions. Chapter 1 introduces IVR and describes a systematic critical review of the use of IVR. IVR is a telephone interviewing technique where the human speaker is replaced by a high quality recorded interactive script to which the respondent provides answers by pressing the keys of a touch-telephone (touchphone). IVR has numerous advantages, including: economy, autonomy, confidentiality, access to certain population groups, improved data quality, standardised interviewing, multi-lingual interfaces, and detailed longitudinal assessments. Despite this, there have been few survey applications of IVR compared to alternative methods such as Computer-Assisted Telephone Interview (CATI). There has not been any evaluation of the use of IVR for asking sensitive questions in surveys or as a tool for health promotion at the community level. A literature review, described in Chapter 2, was conducted to identify other methods of asking sensitive questions. The literature review identified 19 different methods. Those methods that were most successful were those that provided the greatest degree of anonymity to the respondent. It was suggested that IVR may be a suitable method for community surveys. As described in Chapter 3, a custom software called Generalized Electronic Interviewing System (GEIS) was developed. This provided both CATI and IVR interviewing modes. As described in Chapter 4, it was found that the response rate obtained using IVR was unacceptably low, and an alternative interviewing method, the Hybrid method was developed. In the Hybrid method the interview was initiated by the interviewer but completed using IVR with GEIS. As described in Chapter 5, the IVR, CATI and Hybrid methods were used to investigate self-reported rates of alcohol and drug consumption within a telephone household survey of 2880 households. The self-report rates were compared to the National Drug Strategy Household Survey (NDSHS). Response rates did not differ significantly between the CATI and Hybrid methods, however the response rate for IVR was significantly less than the other methods. The Hybrid and IVR methods obtained significantly higher self-report consumption rates for alcohol and marijuana, and significantly higher hazardous drinking scores using Alcohol Use Disorders Identification Test (AUDIT). In Chapter 6 a pilot of an IVR cervical screening brief advice interface is described. A total of 5000 households were contacted by the IVR system. The system randomly selected an eligible woman aged 18-69 per household and determined her cervical screening status. A total of 661 women listened to the IVR message. The IVR call was shown to be acceptable and inexpensive compared to a mail pamphlet intervention. In Chapter 7 a randomized controlled trial of an IVR cervical screening brief advice involving 17,008 households is described. Cervical screening rate data were obtained from the Health Insurance Commission (HIC) for a period spanning six months before and following the intervention. The cervical screening rate was increased in the intervention postcodes by 0.43% compared to the control postcodes, and the increase was greater for older women at 1.34%. This was a desirable outcome since this group is considered to be an at-risk group. The overall conclusion was that IVR technology could be feasibly used to contact women to deliver brief interventions aimed at increasing cervical screening rates, but the cost per screen was likely to be high. It is suggested that an IVR system could be linked to cervical screening registers to more directly and economically contact women, and provide an efficacious complementary approach to the existing letter reminder system.
PhD Doctorate

The research described in this thesis reviewed previous uses of Interactive Voice Response (IVR), developed appropriate software, and employed IVR to obtain self-report of sensitive issues in surveys and conduct brief public health interventions. Chapter 1 introduces IVR and describes a systematic critical review of the use of IVR. IVR is a telephone interviewing technique where the human speaker is replaced by a high quality recorded interactive script to which the respondent provides answers by pressing the keys of a touch-telephone (touchphone). IVR has numerous advantages, including: economy, autonomy, confidentiality, access to certain population groups, improved data quality, standardised interviewing, multi-lingual interfaces, and detailed longitudinal assessments. Despite this, there have been few survey applications of IVR compared to alternative methods such as Computer-Assisted Telephone Interview (CATI). There has not been any evaluation of the use of IVR for asking sensitive questions in surveys or as a tool for health promotion at the community level. A literature review, described in Chapter 2, was conducted to identify other methods of asking sensitive questions. The literature review identified 19 different methods. Those methods that were most successful were those that provided the greatest degree of anonymity to the respondent. It was suggested that IVR may be a suitable method for community surveys. As described in Chapter 3, a custom software called Generalized Electronic Interviewing System (GEIS) was developed. This provided both CATI and IVR interviewing modes. As described in Chapter 4, it was found that the response rate obtained using IVR was unacceptably low, and an alternative interviewing method, the Hybrid method was developed. In the Hybrid method the interview was initiated by the interviewer but completed using IVR with GEIS. As described in Chapter 5, the IVR, CATI and Hybrid methods were used to investigate self-reported rates of alcohol and drug consumption within a telephone household survey of 2880 households. The self-report rates were compared to the National Drug Strategy Household Survey (NDSHS). Response rates did not differ significantly between the CATI and Hybrid methods, however the response rate for IVR was significantly less than the other methods. The Hybrid and IVR methods obtained significantly higher self-report consumption rates for alcohol and marijuana, and significantly higher hazardous drinking scores using Alcohol Use Disorders Identification Test (AUDIT). In Chapter 6 a pilot of an IVR cervical screening brief advice interface is described. A total of 5000 households were contacted by the IVR system. The system randomly selected an eligible woman aged 18-69 per household and determined her cervical screening status. A total of 661 women listened to the IVR message. The IVR call was shown to be acceptable and inexpensive compared to a mail pamphlet intervention. In Chapter 7 a randomized controlled trial of an IVR cervical screening brief advice involving 17,008 households is described. Cervical screening rate data were obtained from the Health Insurance Commission (HIC) for a period spanning six months before and following the intervention. The cervical screening rate was increased in the intervention postcodes by 0.43% compared to the control postcodes, and the increase was greater for older women at 1.34%. This was a desirable outcome since this group is considered to be an at-risk group. The overall conclusion was that IVR technology could be feasibly used to contact women to deliver brief interventions aimed at increasing cervical screening rates, but the cost per screen was likely to be high. It is suggested that an IVR system could be linked to cervical screening registers to more directly and economically contact women, and provide an efficacious complementary approach to the existing letter reminder system.
PhD Doctorate

Diploma thesis deals with the process of validization of Marlowe-Crowne social desirability scale in the Czech context. Despite the fact that social desirability represents an error, which can disrupt validity of obtained data, there is no official Czech version of Marlowe-Crowne scale that could be used in surveys. The aim of this thesis is to create Czech translation of the scale and test its functionality. The Marlowe-Crowne scale is introduced from a theoretical point of view firstly, is discussed its creation, characteristics, its importance in relation to social desirability, critics and problems and possibilities and difficulties of its use in the new context. Further are discussed different ways of translation of research instruments, which could help to preserve their purpose, and method of cognitive interviews as a way of testing of new or adapted research instruments. Modified parallel translation of three translators was used for creation of Czech version of Marlowe-Crowne scale. Cognitive interview, based mostly on verbal protocols, were used for testing of finally translated scale to find out, if scale is applicable in the Czech context without any problems. Interviews were conducted with sixteen respondents with different sociodemographic characteristics. Thesis concludes that scale...

Realizacja sondażu zastawia na badaczy wiele pułapek. Te najbardziej niebezpieczne wiążą się z ryzykiem kierunkowego, systematycznego obciążenia pomiaru. W badaniach poruszających problematykę wrażliwą, pula potencjalnych błędów nielosowych powiększa się o efekt społecznej poprawności, uderzający w dwójnasób w interaktywne techniki wywiadu. Mechanizm zniekształcania odpowiedzi przez zmienną społecznej aprobaty wydaje się dość intuicyjny. Zdarza się, że pytanie sondażu dotyczy kwestii regulowanej społecznie – jego treść, poszczególne warianty kafeterii, odnoszą się do cech i zachowań pożądanych lub takich, które zwykło się uważać za niewłaściwe (a nawet prawnie zakazane). W sytuacji, gdy faktyczna postawa ankietowanego kłóci się z ogólnie przyjmowanymi normami, staje on w obliczu swoistego konfliktu. Z jednej strony poczucie obowiązku nakazuje mu udzielić szczerej odpowiedzi (potrzeba bycia dobrym respondentem), z drugiej – presja wartości skłania ku deklaracji, która nie grozi uszczerbkiem na wizerunku (potrzeba bezpieczeństwa, akceptacji). Normatywna jednoznaczność pytań drażliwych sprawia, że problematyka wyrażona w nich explicite schodzi na dalszy plan – z punktu widzenia osoby badanej są one pytaniami o to, czy łamie ona ustalone zasady lub działa w sposób obyczajowo nieatrakcyjny. Obawa o utratę twarzy powoduje, iż pokusa zafałszowania rzeczywistości bierze często górę, prowadząc do obciążenia odpowiedzi czynnikiem społeczno-kulturowym – znaczącego niedoszacowania (underreporting) i przeszacowania (overreporting) niektórych deklaracji. Nie ulega wątpliwości, iż nawet najstaranniej (i najprzebieglej) zaaranżowane klasyczne procedury nie chronią respondenta przed deficytem prywatności – silnie uspołeczniony kontekst wywiadu wywiera na nim presję, zwiększając prawdopodobieństwo reakcji obronnych. W obliczu tak dużych ograniczeń kwestią czasu wydawało się wypracowanie rozwiązań alternatywnych wobec pytań bezpośrednich – technik, które w sposób definitywny eliminowałyby obciążenia psychologiczne. Wydaje się, że milowego kroku w tym kierunku dokonał S. Warner. W latach 60. minionego stulecia, w krótkim, niepozornym artykule, opisał on nowatorską technikę probabilistyczną (randomized response technique), zapoczątkowując nową epokę w metodologii zgłębiania problematyki zagrażającej – etap intensywnego wzrostu strategii niebezpośrednich. Mimo upływu lat prace nad ulepszaniem algorytmów obliczeniowych, podnoszeniem skuteczności i redukcją przypadkowej wariancji wyników nie ustają, a kolejni autorzy (często przedstawiciele nauk ścisłych), wymieniają się spostrzeżeniami na temat coraz to nowych podejść pomiarowych. W niniejszej pracy postawiono sobie dwa cele. Pierwszym, czysto teoretycznym, było usystematyzowanie procedur zorientowanych na ocenę wielkości i minimalizowanie efektu społecznej poprawności, natomiast drugim, praktycznym – analiza działania randomized response technique. Materiał empiryczny pochodzi z sześciu badań przeglądowych Fundacji Centrum Badania Opinii Społecznej (CBOS). Pomiar przeprowadzano na całej lub części losowej próby dorosłych mieszkańców Polski, zwracając się do respondentów z prośbą o wyjawienie jednej z dwóch ekstremalnie osobistych informacji. Ocena działania technik pomiarowych nigdy nie dokonuje się w próżni – w fazie projektowania procedur konieczne jest ich zmaterializowanie, wypełnienie ram metody treścią. W niniejszym projekcie fabułą uczyniono, intersującą samą w sobie, problematykę doświadczeń aborcyjnych i niewierności partnerskiej. Dość niecodzienny jest też sposób opracowania danych. Typowa dla badań sondażowych obróbka statystyczna (analiza danych numerycznych) nie odgrywa w nim pierwszorzędnej roli. Równie dużo, jeśli nie więcej miejsca, poświęcono bowiem analizom stricte jakościowym (analiza materiału audio). Interpretatywny wymiar prac badawczych przybrał niespotykaną wcześniej skalę. Dzięki triangulacji wyciągane wnioski odcinają się od domniemań – zamiast tego rozprawa przedstawia pogłębione spojrzenie na testowaną strategię badawczą. Postępowanie badawcze wykazało poprawę trafności pomiaru i rzetelność procedury probabilistycznej, ujawniając jednak obecność problemów prowadzących do zaniżenia i lekkiego „przesunięcia” szacunków. Ich wyeliminowanie jest warunkiem koniecznym do prawidłowej akulturacji techniki.
It is said, that when someone breaks the rules, one of the least likely things for him is to admit doing it to a stranger. If you break the law or behave in the way that may be perceived as inappropriate or unattractive, you will probably not tell the truth. The reason why you lie is the psychological distress – the need for social approval, which arises as a consequence of the pressure of cultural and social norms. In surveys this kind of distortion is called the social desirability effect, and it is thought to be one of the most common systematic biases that must be faced by researchers who perform standardized interviews. The thesis has been an attempt to describe processes underlying the social desirability effect and systematize techniques designed to estimate and minimize this bias. In addition to these theoretical purposes the author has examined one indirect method for eliciting sensitive information – the randomized response technique. The empirical part of the dissertation was based on results of six surveys. Each of them was conducted on the large nationally representative random sample of Polish adults and carried out by the Public Opinion Research Center (CBOS). The interests of researches were extremely personal. Thus, in two of them respondents (men and women) were asked if they had ever been unfaithful to their spouse or partner. In four others the women were asked if they had ever had an abortion. Both problems were measured in traditional and indirect way. The comparison of quantitative results combined with the qualitative analysis of the interviewer-interviewee interaction made it possible to assess the accuracy and the efficiency of the randomized response technique. The investigation has shown that despite higher level of complication, the indirect procedure increases the psychological comfort of respondents encouraging them to reveal shameful experiences. Significant growth of estimates goes with their repeatability. Furthermore, interviewees are rather aware of their role and understand the task in which they are involved. All of these findings let the social scientists be quite optimistic about the future of the measurement of sensitive issues.