1 PDF of the PPT

2022-09-25 1 TAKING UP THE QUIZ! WHICH OF THE FOLLOWING IS TRUE ABOUT FEATURES OF A QUASI-EXPERIMENTAL RESEARCH DESIGN? • Manipulation, control group, randomization • Use of correlational approach • Manipulation, but no control group or randomization • No manipulation of independent variable • Quasi=resembling • This type of experiment lacks randomization, but the independent variable is still manipulated before the dependent variable is measured. • The first possible answer is incorrect as it includes randomization. The second answer is a different kind of study. The fourth is incorrect because with quasi-experimental there is manipulation of the independent variable. THE SUM OF THE VALUES OF A VARIABLE FOR A SET OF OBSERVATIONS, DIVIDED BY THE NUMBER OF THE OBSERVATIONS IN THE SET REFERS TO: • Variance • Standard deviation • Mean • Mode • For this question, probably easiest to just look up all the terms: • Variance, or dispersion, roughly refers to the degree of scatter or variability among a collection of observations. https://methods.sagepub.com/reference/enc yclopedia-of-survey-research-methods/n620.xml • Standard deviation: how much the answers spread from the mean – can be tightly (i.e. most people answered or measured almost the same) or can be wide (extremes from the mean) • Mean = the average, sum of the data divided by the number of entries • Mode = the value that appears most often A RESEARCHER WANTS TO STUDY THE EFFECTS OF A MAN DEVELOPING BREAST CANCER ON HIS PARTNER’S (WIFE OR SAME SEX PARTNER) DECISION TO WORK OUTSIDE OF THE HOME. WHICH OF THE FOLLOWING FACTORS NEGATIVELY AFFECTS THE FEASIBILITY OF THIS STUDY? • The local region has a high percentage of divorce • Breast cancer in males represented less than 1% of cases in the region during the year previous to the study • The partners (wife or same sex partner) may not have working outside the the home prior to their husband’s diagnosis of breast cancer • A comprehensive review of the literature shows no studies that examined this question were conducted during the past 15 years • This question is to prompt that there are some studies that will likely never occur as the thing being studied might be too rare, or the question (like this one) too specific. The rate of divorce is not what the question is looking at, whether the spouse worked prior to the diagnosis doesn’t mean we can’t study it post diagnosis, and if the lit review showed no studies that usually means we should research it, rather than decrease the feasibility. • This question/study could be broadened to include “all cancer” or a more common cancer in males (i.e. testicular), which would make the data more available and therefore more feasible. WHICH OF THE FOLLOWING TECHNIQUES OR ACTIONS REPRESENTS CONSTANCY IN DATA COLLECTION FOR CONTROL OF EXTRANEOUS VARIABLES IN A STUDY EXAMINING THE EFFECT OF HORMONE REPLACEMENT THERAPY ON BONE DENSITY LOSS IN POSTMENOPAUSAL WOMEN? • Drawing blood from all subjects at the same time of day • Always having the same nurse or phlebotomist always draw blood from all subjects • Ensuring that all subjects are postmenopausal women taking hormone replacement therapy • Using the same room for drawing blood from all subjects • In this question, it is best to know what constancy means before you go at it – this is what LoBiondo-Wood & Haber (2013) describe as the cookbook like recipe – same conditions, timing of the data, instruments and procedures. Looking at the answers, the two middle ones are not part of the “cookbook” (the nurse or phlebotomist is the chef) and the subjects are not part of the constancy either (they may be included/excluded at the set up of the study). The time of day and same room could both be argued to be potential, but the “most right” answer for a blood draw would the time of day, since hormones etc. vary throughout the day. The room would not likely impact the blood draw, but in other experiments that could matter. Keeping the room the same would also limit the study participants to one town, which would reduce the strength of the study. WHICH STATEMENT ABOUT CHI-SQUARE TEST ( X 2 ) IS THE MOST ACCURATE? • Chi square is a statistical test that is used to assess differences in proportions of categorial data • A Chi Square test is used to determine the relative frequency of an event in repeated trials under similar conditions • A Chi Square test determines whether there is no difference between the groups in a study, or no association between the variables under study. • A Chi Square test refers to tolls used when data are collected at the ordinal level. • The definition from LoBiondo-Wood & Haber (2013) is: “a nonparametric statistic used to determine whether the frequency found in each category is different from the frequency that would be expected by chance…if the calculated chi-square is high enough, the researcher would conclude that the frequencies found would not be based on chance alone, and the null hypothesis would be rejected” page 370 • So the most accurate answer would not be about relative frequency with repeat trials; would not be about differences between groups and would not be about data at an ordinal level (or not exclusively)

2022-09-25 2 FROM THE FOLLOWING EXCERPT DETERMINE THE INDEPENDENT VARIABLE: "AT 4 MONTHS, THE MEAN WEIGHT CHANGE IN THE [EDUCATIONAL] INTERVENTION GROUP WAS .81 LBS, WHILE THE MEAN WEIGHT CHANGE IN THE STANDARD CARE GROUP WAS 7.17 LBS" (LITTRELL, HILLINESS, KIRSHNER, PETTY, & JOHNSON, 2003, P. 240) • time frame • intervention = nominal weight = ratio • educational intervention • weight gain • The independent variable is the one that is manipulated by the researcher. The only answer above that the researcher can impact is the educational intervention. THE PROBABILITY OF MAKING A TYPE II ERROR IN ANY STATISTICAL TEST IS TERMED AS: • Alpha • Phi • Delta • beta • LoBiondo-Wood & Haber (2013) page 364 – “type II error – aka beta” WHICH OF THE FOLLOWING FACTORS IS CONSIDERED A MEASURE TO KEEP THE CONDITIONS OF A QUANTITATIVE STUDY UNIFORM? • Methods • Accuracy • Control • Validity • This question is another where we just look to define the answers • Methods is the broad term to catch all possible ways of research; • Accuracy LoBiondo-Wood & Haber (2013) page 201 all aspects systematically and logically follow from the research problem. Pilot study is a way to test the accuracy of the design before the larger study is undertaken (a test run); • Control is holding the conditions uniform (page 200); • Validity whether a measurement instrument actually measures what it is purported to measure (i.e. in one episode of the office, Pam eggs Dwight into running around the building as fast as possible, and she will time him. She admits to using a thermometer. This study of Dwight’s speed is not using a valid instrument.) page 580 THE DEGREE TO WHICH A DATA COLLECTION TOOLS ACCURATELY MEASURES WHAT IT INTENDED TO MEASURE IS KNOWN AS THE: • Validity • Dependability • Credibility • reliability • This was described in the last slide in the Pam and Dwight example. Reliability is about consistency and predictability, the same results on repeated measures. A standard ruler is reliable as it will always tell you the length of an object, “eyeballing” the length is unreliable as you might not guess the same length each time even if the item is the same length (i.e. if it was near something smaller or larger, you may guess it different despite being the same size; and your guess is not predictable to other people). • Dependability is not really anything to do with research, unless it is about your team member! • Credibility is a characteristic in qualitative research that refers to the accuracy, validity and soundness of the data. LoBiondo-Wood & Haber (2013) p 325 These lines are all the same length, so if you used a standard ruler you would get the same answer. But if you guessed, you would like be unreliable. WHICH OF THE FOLLOWING SITUATIONS MEETS A CRITERION FOR INFERRING CAUSALITY BETWEEN PERFORMANCE OF EXERCISE AND ADAPTATION BEHAVIOUR OF WOMEN RECEIVING CHEMOTHERAPY FOR BREAST CANCER? • No differences in adaptation behaviours were observed between subjects in the experimental group and subjects in the control group • The level of adaptation behaviours increased proportionately with increase in the intensity of the exercise • Adaptive behaviour occurred in the control group over time • Adaptation behaviour began before the exercise program was initiated. • Causality – inferring causality requires three criteria to be met: • Causal variable and effect variable must be associated with each other • The cause must precede the effect • The relationship must not be explainable by another variable • The first answer really has nothing to do with causality, the 2 nd answer is about the increasing of both the variables (so if the exercise was introduced first, and then we measured the adaptive behaviours, that would be potential causality); the 3 rd one has nothing to do with causality and for the 4 th if the adaptive behavior began before the exercise, there would not be causality. IN A STUDY, RESEARCHERS WANTED TO LEARN THE IMPACT OF TELEVISIONVIEWING ON GRADES. RESEARCHERS DISCOVERED THAT AS THE NUMBER OF HOURS SPENT WATCHING TELEVISION INCREASED, GRADES DECREASED. THIS STUDY IS AN EXAMPLE OF WHAT TYPE OF RESEARCH METHOD? • experimental method • correlational research • case study • naturalistic observation • In this question it is easier to think through what each one means – this is not an experiment, as the researchers are not manipulating the tv (they could do one this way, but it is not what they are doing here); it is not a case study of a few interesting findings; naturalistic is a qualitative type of research where a researcher goes to a natural setting to observe phenomenon (and we aren’t learning about qualitative this course) – so the answer left is correlational, which makes sense as this is a type of nonexperimental research looking at the relationship between two variables

2022-09-25 4 WHICH OF THE FOLLOWING WOULD BE USED TO ASSESS WHETHER MENTAL HEALTH PATIENTS WHO RECEIVED COGNITIVE BEHAVIOURALTHERAPY VIA SMART PHONES REMAIN OUT OF HOSPITALS FOR LONGER PERIODS OF TIME. • predictive, statistical survey • cross-sectional survey • causative, correlational survey • longitudinal, comparative survey • This one might be tackled through elimination we know its not a causative/correlational survey as we are not looking for some variable causing an event; it is not a cross sectional survey either. We are comparing over time, a variable that some people will get and some will not get, the CBT via smart phones but the researchers are not necessarily influencing/introducing the CBT, just looking at two groups WHICH OF THE FOLLOWING IS NOT RELATED TO DEALING WITH EXTRANEOUS VARIABLES IN A QUANTITATIVE RESEARCH? • Homogeneity • Randomization • repeated measures • Bracketing • LoBiondo-Wood & Haber (2013)page 203 Extraneous variables interfere with the operations of the phenomena being studied (age, gender) and ways of controlling include homogeneous sample (all the same), consistent data collection procedures, manipulation of the independent variable and randomization. • Bracketing is a qualitative research term, defined on page 160 of LoBiondo-Wood & Haber (2013) WHY SHOULD A RESEARCHER AVOID DRAWING CONCLUSIONS OR MAKING GENERALIZATIONS BASED ON THE EXPERIENCES OF A SMALL NUMBER OF PARTICIPANTS? • Small samples invalidate hypotheses • The researcher may be unable to eliminate his or her bias • Data obtained from a small number may not adequately represent the phenomenon • Small samples increase historical threats to internal validity. • Small samples would not invalidate a hypothesis, small samples would not make a researchers bias any easier/harder to eliminate and small samples would not increase threat to internal validity; but I think we can all agree from our discussion, a small sample might not be enough to really demonstrate the phenomena that were are studying. WHY WOULD A NURSE RESEARCHER OPT FOR A QUASIEXPERIMENTAL STUDY DESIGN INSTEAD OF AN EXPERIMENTAL STUDY DESIGN? • The researcher has not conducted research before • The study is planned to be conducted in a laboratory setting. • An experimental treatment is not part of the study. • Full experimentational control is not possible • This one is fairly straightforward – we would opt for a quasi-experiemental when we could not randomize something or have an RCT type study done for many of the reasons we discussed these last weeks. It would have nothing to do with never having researched before, have nothing to do with a lab or natural setting and if the experimental treatment is not part of a study, I guess that means it is not a study! A STANDARD DEVIATION IS: • Abbreviation for the number of measures in a given data set (the measures may be individual people or some smaller pieces of data, such as blood pressure readings) • the values that occur most frequently in a data set • 68% of the values in a normal distribution fall within + 1 of this value • an old abbreviation for the mean • Standard deviation LoBiondo-Wood & Haber (2013) 579 – a measure of variability, measure of average deviation of score from the mean. More detail is available on page 356, and this might be helpful to review https://www.youtube.com/watch?time_con tinue=116&v=hQTvdD8vtio&feature=emb _logo WHICH OF THE FOLLOWING THREATS TO INTERNAL VALIDITY IS MORE LIKELY TO OCCUR WITH A LONGITUDINAL DESIGN? • history effects • mortality effects • instrumentation effects • selection bias • We know that longitudinal studies (likes NHS or GUTS) have been going on for years. So, people will die, but it is also about general attrition from the first to the last data collection point. Selection bias is not more likely in a longitudinal study, history effects/threat – events happening outside the experimental study (the example on page 207 LoBiondo-Wood & Haber (2013) about breastfeeding education the history threat are the commercials on tv and newspaper

2022-09-25 5 WHICH OF THE FOLLOWING TYPES OF QUANTITATIVE RESEARCH REQUIRES THERE TO BE A CONTROL GROUP AND AN EXPERIMENTAL GROUP? • pilot study • experimental design • surveys • secondary data analysis • To have a control and an experimental group means we have an experimental design, and the other options are obviously not correct if we used process of elimination. WHICH OF THE FOLLOWING STATEMENTS ABOUT INCLUSION AND EXCLUSION CRITERIA IS TRUE? • Inclusion criteria are used to define dependent variables, and exclusion criteria are used to define independent variables • Inclusion criteria are used to define independent variables, and exclusion criteria are used to define dependent variables. • These criteria are not needed when the sample being studied is composed of events. • These criteria serve to ensure that the characteristics of the sample and the population are congruent • we have inclusion and exclusion criteria to gather a sample that will be congruent, this is explained on page 260 of LoBiondo-Wood & Haber (2013). WHICH OF THE FOLLOWING STUDY TITLES SUGGESTS A CROSS-SECTIONAL DESIGN? • Effect of Prenatal Parenting Classes on Maternal– Infant Bonding in the Early Postpartum Period • Change in Self-Esteem Over Time Among Women Participating in a Weight Loss Support Group • Relationship Between Self-Esteem and Successful Breastfeeding at 1 Month and 6 Months After Delivery • Women's Appraisal of the Diagnosis Within the First 48 Hours After Initial Breast Cancer Diagnosis • Cross sectional study – nonexperimental research in which data at one point in time – the immediate present- are examined. So we are looking for a “snapshot” of a life – not the classes they took, not the change over time when in a weight loss group, no self esteem and breastfeeding at 1 and 6 months so the “snapshot” is the Women’s appraisal of diagnosis within the first 48 hours (the snapshot). IN A STUDY CONDUCTED AT A LARGE LONG-TERM CARE FACILITY, TWO DATA COLLECTORS EXAMINED 56 PRESSURE ULCERS ON 40 DIFFERENT SUBJECTS. THE EXAMINATIONS WERE INDEPENDENTLY PERFORMED BUT ON THE SAME DAY. A COMPARISON OF THE RESULTS INDICATED THAT THE DATA COLLECTORS, WHO USED THE BRADEN SCALE FOR PRESSURE ULCER RISK ASSESSMENT, HAD IDENTICAL SCORES FOR 39 OF THE 40 PATIENTS. WHAT CAN BE DETERMINED FROM THIS FINDING?" . WHAT CAN BE DETERMINED FROM THIS FINDING? • Interrater reliability between the two data collectors was high. • The data collection method was inappropriate for the phenomenon under investigation. • In order to establish interrater reliability, both data collectors should have examined each pressure ulcer at the same time • Interrater reliability between the two data collectors was low • Two different data collectors had the same score in all but one person, so there was good inter-rater reliability between the two, so it was high and not low. Examining the pressure ulcer at the same time would not really impact the scoring (unless it was by days) for the Braden scale, and when I wrote it originally it was confusing as it was talking about the ulcers rather than the prevention. Good catch! USING A NEW TESTING INSTRUMENT TO MEASURE BEHAVIOR, THE FIRST TIME A PARTICIPANT TAKES THE TEST,THEY RECEIVE A SCORE OF 25. THE SECOND TIME THEY TAKE THE TEST, THEY SCORE A 12. THE THIRD TIME THE TEST IS TAKEN THE PARTICIPANT RECEIVES A SCORE OF 38. BASED ON THE SCORES, THE NEW TESTING INSTRUMENT HAS • high reliability • low reliability • high validity • low validity • So this one is easy to do process of elimination. The scattered results mean it is going to be “low” something right? Validity is the measurement tool measuring what it says it does; and the reliability is consistency or constancy of a measuring instrument, getting the same results on repeat measures. So, it would have low reliability, since it had different answers each time. THE PURPOSE OF RANDOM SAMPLING IS TO ENSURE • representation of specific subgroups of the population • representativeness of the sample • sufficient sample size • This one should be straightforward by process of elimination – the sample size has nothing to do with the random sampling, it will not necessarily be a good representation of the specific subgroups (Depending on what the source of the list to be randomized is) and so it really is to get a representation of the sample – i.e. random sampling of names off a list of those who just renewed their health card in the last month to get feedback on the new system during COVID. It should be a representative sample of the user population.