Evidence-Based Practice

Statistics in Social Work The steps of evidence-based practice include formulating an answerable question. How does knowing about statistics improve our ability to be an evidence-based practitioner at this step? How understanding statistical principles can enable you to better understand if a question is answerable or not. Are 'baselines' in descriptive function, or predictive function available for assessment. In application of statistics to social phenomena, the frequency, duration and intensity of the subject tested will contribute to analysis where more than nominal (i.e. numbered) distributions are involved.

Merely 'counting' a population is not a significant activity in statistical renderings as independent variables require dependent variables in order to acquire statistical meaning. Evidence-based practice references studies that 'replicate' existing tests, toward reinterpretation of former statistical outcomes in a new study of parallel significance, with variables of the same classification. Patterns in longitudinal tests over time offer insights into stasis or transformations in variables, and potentially contributory elements. Example: census analysis of populations looks at 'number' of people, of 'xyz' ethnic classification with an 'income' of a certain tier.

Aggregate outcomes have 'meaning' where discretionary categorization is considered as part of a binomial distribution for significance beyond normal distributions (i.e. mean). Comparison of data from two census indexes offers longitudinal evidence to future practice. b. How understanding the basic principles of hypothesis testing can help you create search terms to apply to your literature. As practitioners, we employ the taxonomy of found in the framework to procedural protocols in our evidence-based practice settings.

Understanding the basic principles of hypothesis testing, where propositions acquire meaning through investigation, furthers the rationale of 'evidence' as a methodology to building best practices into future research and in the agency setting. Informed hypotheses are derived from prior knowledge of existing research and reported outcomes. Outcomes are significant in terms of evidence, and the knowledge sharing capacity is increased through the feedback loop of 'replication' and recommendation .Cited terms within those professional investigations serve as lexicon to policy and procedure. 2.

The steps of evidence-based practice also include locating relevant evidence to answer your question. How can knowing about statistics assist you in sorting through the available evidence for the best possibilities of answering your question? a. How knowing about statistics can help you quickly identify, from a study abstract, whether a particular piece of evidence is appropriate to your question. It is said that 'some statistics are better than others.' Quality reporting on outcomes to primary research must still be reviewed for sufficient assumptions to the initiation of tests.

For example, the interpretation of population data as findings through statistical measurement in t-tests, standard deviations, effect size and mean, will offer indicators to large sample size populations, where aggregate and other relatively abstract occurrences have taken place. b. How knowing about statistics can help you identify, from a study abstract, whether appropriate statistics were used in the study. Equidistance between two points as seen in standard deviations, narrates little where a broad distribution is not merely intended to show stratified events or classes of integers.

An example of where assumption to the type of test employed is a standard deviation recording, would be income. In this case, deviations in income are proven significant. In a study focused on variables rather than integers, standard deviations will have little to no significance. Decision making models where more than two variables present would be found to have zero standard deviation where one outcome is not hierarchical in distribution over the other. 3. The steps of evidence-based practice require you to critically evaluate evidence.

How does knowing the underlying principles of hypothesis testing, covariance and probability assist you in your critical evaluation? a. How assumptions to various statistics can help you to assess the accuracy of the results of the study. Assumptions to statistics require foresight in design of a statistical model for research testing. Well thought out investigations assess correlation of hypothesis to subjects through properties of measurement. Indeed, the analytical process to statistical testing is already present within the logic model of the research design.

Delineation between variables, constants, and the dependent and independent factors to those data, will also assist in discerning predictor or known value (I.V.), and criterion or what is predicted (D.V.) in formation of the study. Accuracy in outcomes may be examined against those initial assumptions to check validity in measure, and in outcomes. Back end testing of additional questions utilizing data from the sample, and test type, performs 'audit' of sorts on the research.

While not typically necessary, as outlandish findings are invariably obvious to professionals whom have been working at practice, in training this applied method of assessment is perhaps the best way to learn how to form a strong hypothesis. b. How using knowledge about effect size can prevent you from applying a relatively weak intervention on the basis of statistical significance alone. Risk Probability is a simple answer here.

Knowledge of effect size will enable discernment where applicability as evidence is not valid, when findings to a small study only reveal insignificant outcomes in terms of statistical significance. In such a case, it is impossible to say that 'universal' or even important recommendation to practice can be made. Preliminary clinical trials fall in this category. 4. The steps of evidence-based practice require you to evaluate outcomes at the individual level. How does knowing about statistics, especially the limitations of statistics, enable you to effectively evaluate case levels outcomes? a.

How statistical approaches to single case data can have limitations. Again, single case statistical analyses do not provide the multi-scale inferences that investigative research offers. Single case data are normally rendered to show a 'trend.' Trends may be operational within a given context that might seem relevant to case management in social work, but trends may also be 'nationwide' studies which have produced an effect size related to family law case decisions.

In such a case, the utility of single case outcomes is informative to practice, but is not a guaranteed probability in the administration of a single client case. Here 'single case data' are not applicable to an individual intervention, because the probability that a family law matter is decided one way or the other, is always a 50/50 chance. b. Ho using your knowledge about effect size applies to assessing the clinical significance of your intervention. In circumstances where clinical testing has been abundant (i.e.

alcohol addiction), certain statistical outcomes to discretionary studies where subjects were drawn from a proximate convenience sample of an almost identical population, and this has been done consecutively in longitudinal analysis for many years, and in a number of sites, certain criteria in part to 'evidence-based practice' where rules in conduct with such clients are considered, are based on fairly meaningful and accurate statistical times of rehabilitation etc.

The sheer magnitude of the 'problem' within the scope of discretionary research dedicated to such analysis is usually a strong indicator that consistency in outcomes is equivalent to reliability.