Norval F. Pohl

Using Retrospective Pre-Ratings to Counteract Response-Shift Confounding

A potential confounding effect present in research designs which employ self-report instruments is response-shift, i.e., a change in the internalized standard by which a subject rates himself. Howard et al. (6) suggest that when self-report measures must be used to measure behavioral changes there should be a modification of Campbell and Stanley’s design (2) so as to include a “retrospective pre-rating.” In the current research, the response-shift phenomenon is demonstrated in the context of a typical classroom setting. Retrospective pre-ratings are shown to yield significantly more accurate estimates of preinstruction knowledge than simple pre-ratings. However, there was no statistically significant difference between correlation co-efficients relating “behavioral change” and “self-reported change” when self-reported change was defined in terms of retrospective pre-ratings rather than in terms of simple pre-ratings.

Scale Considerations in Using Vague Quantifiers.

The responses to many survey questions are often made in terms of vague quantifiers such as very often or occasionally. However, we know little about the scale characteristics of such quantifiying terms. This paper reviews the literature on the use of frequency expressions as questionnaire anchor points. Data are presented which suggest that the popularly used five-point response category set of always, often, occasionally, seldom, and never is not intervally scaled. Utilizing data obtained through ratio estimation, an approximately equal interval measurement scale is developed for a five-point response category set.

Student Achievement in Computer Programming: Lecture vs Computer-Aided Instruction.

This paper discusses a study of the differences in student learning achievement, as measured by four different types of common performance evaluation techniques, in a college-level computer programming course under three teaching/learning environments: lecture, computer-aided instruction, and lecture supplemented with computer-aided instruction.Three samples of students were obtained and matched on relevant variables. The analysis of variance, matched groups design, detected significant differences among treatment groups (teaching/learning environments) on two of the four performance evaluation techniques.

Computer-Assisted Instruction Augmented with Planned Teacher/Student Contacts

This paper discussed a study of the differences in student learning achievement and retention in a college-level statistics course taught in a variety of teaching/learning environments. Statistical test results revealed that students experiencing a CAI environment performed no differently on achievement or retention tests than students experiencing a traditional lecture/discussion environment. However, students experiencing an “enriched” CAI environment (CAI plus planned teacher/student contacts) performed significantly better on achievement tests than students experiencing any of several other environments, including: lecture/discussion, lecture/discussion supplemented with planned teacher/student contacts, PI texts, PI texts supplemented with planned teacher/student contacts, and CAI.

Reducing Item-Specific Nonresponse Bias.

This article describes a method of dealing with partially-responded-to questionnaires in survey research which utilizes nonparametric Bayesian discriminant analysis to predict missing responses based on “profiles” of the partial and the “complete” respondents. Some background of nonresponse bias in survey research is discussed, and two other methods of dealing with partial nonresponse bias are briefly criticized. The Bayesian method is detailed along with its statistic, and limitations on the use and effectiveness of all the methods are considered.

Evaluating Franchise Investment Opportunities using Probability-Tree Analysis

A franchise investment opportunity is normally evaluated by the prospective investor in terms of financial ratio analysis, simple break-even analysis, and cash and capital budget analysis of average sales and profit figures or ranges provided by the franchising organization. In addition to this basic evaluation of the venture, a prospective investor can, using a probability-tree analysis, take into account the probabilities of occurrence of various profit levels based on his subjective judgment ratings of proposed locations and of proposed managerial ability. Such an analysis allows the franchise investor to gain maximum decision-making information from available sales and profit data.

Random Response Techniques for Reducing Non-sampling error in Interview Survey Research.

For “sensitive-area” questions in interviewing, random response techniques can be useful in establishing the types of questions to be asked and the methods by which the respondent can confidentially provide answers. This paper discusses three randomized response techniques (the dichotomized question, the unrelated question, and the single question/random answer), the simple probability theory that makes the techniques work, and some questions about the practical application of these techniques by the educator.

The Relative Classificatory Ability of the Linear Discriminant Function and the Bayesian Taxonomic Procedure.

AbstractThe purpose of this study was to compare the relative classificatory ability of the Linear Discriminant Function (LDF) and the Bayesian Taxonomic Procedure (BTP) when these techniques are applied to multivariate normal and nonnormal data with differing degrees of overlap in the distributions of the predictor variables. The findings and test results led to the conclusion that the LDF is an extremely robust classificatory technique. In general, the LDF should be used in solving discriminatory problems involving continuous univariate data. Only when the data are extremely skewed should a researcher consider using the BTP in preference to the LDF.