Social Work Research Questions and Answers

Answer: B

Failing to reject a false null hypothesis is a Type II error.

A Type II error (probability beta) occurs when the null hypothesis is actually false but the test fails to reject it, so a real effect or difference is missed. Its probability falls as sample size, effect size or the significance level rises; the power of a test is \(1 - \beta\).

A Type I error is the opposite: rejecting a null hypothesis that is true, with probability alpha. Sampling error is the chance difference between a sample statistic and the population value, and non-sampling errors arise from measurement, non-response or processing; neither is defined by a wrong decision about the null hypothesis.

Answer: B

The preferred design is the single subject research design.

Single subject (single system) designs measure a client's target problem repeatedly during a baseline phase (A) and then during the intervention phase (B), as in AB, ABA or multiple-baseline designs. Because they can be built into everyday practice with one client, family or group, social work research texts recommend them for evaluating the effectiveness of one's own interventions.

Descriptive and exploratory designs describe or explore a situation and do not test whether an intervention worked. Group experimental designs can test effectiveness, but they need control groups and random assignment that are rarely practical for a worker evaluating his or her own practice.

Answer: D

The non-probability method is purposive sampling.

In purposive (judgement) sampling the researcher deliberately picks cases thought to be typical or informative. Because selection depends on judgement, not every member of the population has a known chance of being chosen, which is what makes it non-probability sampling.

Simple random, stratified and cluster sampling are all probability methods: every unit, or every cluster, has a known, non-zero chance of selection, chosen by a random process within the population, the strata or the clusters.

Answer: B

It is a complex hypothesis.

A simple hypothesis predicts a relationship between one independent and one dependent variable. A complex hypothesis predicts the relationship between two or more independent variables and two or more dependent variables, for example that age and income together affect both health and life satisfaction.

A statistical hypothesis is a statement about population parameters tested with statistics. Directional and non-directional hypotheses differ in whether they predict the direction of a relationship, not in how many variables they involve.

Answer: B

The Social Distance Scale was developed by Bogardus.

Emory S. Bogardus devised the scale in the 1920s to measure people's willingness to accept members of other social groups at increasing levels of closeness, from admitting them to one's country, through having them as neighbours and colleagues, to admitting them to close kinship by marriage. It is a cumulative scale: accepting a closer relationship implies accepting the more distant ones.

L. L. Thurstone developed the method of equal-appearing intervals and Rensis Likert the summated rating scale. Wellington is not associated with any attitude scale.

Answer: B

The key is Both (A) and (R) are true and (R) is not the correct explanation of (A).

The Assertion is true: quantitative researchers prefer probability sampling because every unit has a known chance of selection, so the results can be generalised to the population with a measurable margin of error. The Reason is true only in the weak sense that non-parametric tests can be used on a probability sample. It is not why the method is preferred. The preference comes from representativeness and the ability to make statistical inferences.

Option A would make the tests the explanation, which is not so. Option C and option D treat the Reason or the Assertion as false.

What could not be confirmed: whether the Reason is best called true or false. The answer follows the University's key.

Answer: A

The universe of study consists of all the elements covered under study.

In research methodology the universe is the entire set of units, such as persons, households or institutions, about which the study seeks to draw conclusions. It is defined by listing what it is made of: every element that falls within the scope of the study.

'Population' is another name for the universe rather than a description of what it consists of. A sample is the part of the universe actually studied, and randomization is a procedure for selecting or assigning units, not a set of elements.

Answer: C

The answer is (A) is true but (R) is not true.

Quantitative researchers do prefer standardised scales, but the reason is their established reliability and validity, uniform administration and scoring, and the availability of norms, which make measurements accurate and comparable across studies.

Being easily available is not why such scales are chosen, and it is often not even true, since many standardised instruments are copyrighted, costly or need permission and training. The Reason is therefore false, which rules out every option that treats it as true.

Answer: D

The test that is not non-parametric is Student's t-test.

The t-test is a parametric test: it compares means and assumes interval or ratio data drawn from normally distributed populations, with equal variances in the usual two-sample form.

The other three are non-parametric tests, which make no such distributional assumptions and often work on ranks. The Kolmogorov-Smirnov test compares distributions, the Wilcoxon signed-rank test compares two related samples, and the Mann-Whitney U test compares two independent samples, the non-parametric counterpart of the independent t-test.

Answer: C

The key's answer is Sequential random sampling.

The researcher starts from a list of patients, a sampling frame, and draws the sample from it on a random basis, so the method must be a random one that works through a prepared list.

The others do not fit. Stratified sampling first divides the population into strata such as age or sex, which the question does not mention. Snowball sampling finds new subjects through referrals from existing ones, and quota sampling fills fixed quotas by the researcher's choice; both are non-random.