Chapter 26: Interpreting Research Outcomes

Chapter 26: Interpreting Research Outcomes

 

Complete Chapter Questions With Answers

 

Sample Questions Are Posted Below

 

MULTIPLE CHOICE

 

1. A researcher studies the effect on reading comprehension level of providing fourth-grade boys who are slow and ponderous readers with illustrated comic books for a six-week summer session instead of textbooks, hypothesizing that reading associated with pictures will results in a higher reading comprehension level. The researcher set a level of significance at p <.05. Analysis of the data from a study indicates that there was a significant difference between the two groups, which were randomly selected and composed of 100 subjects each. However, the experimental comic book group’s scores improved less than did the scores of the boys who used the textbooks. The measured p-level was 0.026. What type of finding is this?

a.	Significant results that are in keeping with those predicted by the researcher
b.	Nonsignificant results
c.	Significant results that oppose those predicted by the researcher
d.	Mixed results

 

 

ANS:  C

Significant results opposite those predicted, if the results are valid, are an important addition to the body of knowledge. An example would be a study in which the researchers proposed that social support and ego strength were positively related. If the study showed that high social support was related to low ego strength, the result would be the opposite of that predicted. Such results, when verified by other studies, indicate that the theory being tested needs modification and refinement. Because these types of studies can affect nursing practice, this information is important.

 

DIF:    Cognitive Level: Application           REF:   Page 594

 

2. The analysis of the data from a study indicates that there was a significant difference between the two groups, which were randomly selected and composed of 100 subjects each. The instruments measuring stress, mood, and various scaled measures of emotion were well normed, widely tested, and suitable. As the researcher begins interpretation of the data, she discovers that during data collection one of her assistants had recorded the demographic item “household size” incorrectly, understanding the item as “number of people who live with you.” There is no way to determine which research assistant worked with which research subject. What effect did this way of recording the data have on the study findings?

a.	That item cannot be analyzed as a ratio item and will have to be treated in a different way.
b.	None, because the numbers will be the same either way the data are collected.
c.	That particular item will have to be discarded, because it can no longer be analyzed.
d.	None, because post hoc statistical analysis will automatically adjust for this type of “error.”

 

 

ANS:  A

Many activities that occur during data collection affect the meaning of study results. Did your study have a high refusal rate for subject participation, or was the attrition high? Was the sample size sufficient? Did strategies for acquiring a sample eliminate important groups whose data would have influenced the results? Did you and your research team achieve intervention fidelity when the treatment was implemented? Did unforeseen events occur during the study that might have changed or had an impact on the data? Were measurement techniques consistent? What impact do inconsistencies have on interpreting results?

Methodological limitations result from factors such as nonrepresentative samples, weak designs, single setting, limited control over treatment (intervention) implementation, instruments with limited reliability and validity, limited control over data collection, and improper use of statistical analyses. Nominal level of measurement is the lowest of the four measurement levels or categories. It is used when data can be organized into categories of a defined property but the categories cannot be ordered. Data that can be measured at the ordinal level can be assigned to categories of an attribute that can be ranked. In interval level of measurement, distances between intervals of the scale are numerically equal. Interval scales are assumed to be a continuum of values. Ratio level of measurement is the highest form of measure and meets all the rules of the lower forms of measures.

 

DIF:    Cognitive Level: Synthesis              REF:   Page 592

 

3. In a cross-sectional study, a researcher studies career trajectories in nurses over the past twenty years. There are five correlational research hypotheses. Why are mixed results, related to significance of findings, to be expected?

a.	The study is cross-sectional, and the economy has made bedside care a more practical decision over the past few years.
b.	Groups across time cannot be expected to have the same responses to career advancement.
c.	The study is merely correlational.
d.	Five statistical hypotheses are being tested; not all of these can be expected to demonstrate statistical significance.

 

 

ANS:  D

Mixed results are probably the most common outcome of studies. In this case, one variable may uphold the characteristics predicted whereas another does not, or two dependent measures of the same variable may show opposite results.

 

DIF:    Cognitive Level: Analysis                REF:   Page 594

 

4. A researcher working for a company that produces a standardized exam that predicts whether or not nursing students will pass the national licensure exam tests this tool at a level of significance of a = .01, with power = .91. The results are statistically significant. What interpretation can the researcher justifiably make?

a.	Findings are important to nursing schools, and valuable for them.
b.	The exam should be used by all nursing schools.
c.	There is good evidence that the results can be replicated, with similar results.
d.	Results will be the same if the study is repeated with another sample.

 

 

ANS:  C

The researcher chooses the probability of making a Type I error when setting α, the level of significance. Nursing researchers usually set a level of significance of p <.05. A level of significance of p <.01 is much more stringent, and decreases the possibility of error to 1%.

Significant results that coincide with the researcher’s predictions are the easiest to explain and, unless weaknesses are present, validate the proposed logical links among the elements of the study. There is good evidence that the reports are reproducible.

 

DIF:    Cognitive Level: Analysis                REF:   Page 594