Help
Options
Didn't know it?
click below
click below
Knew it?
click below
click below
Don't Know
Remaining cards (0)
Know
retry
shuffle
restart
0:00
Embed Code - If you would like this activity on your web page, copy the script below and paste it into your web page.
Normal Size Small Size show me how
Normal Size Small Size show me how
Research MidTerm
Clinical Research Mid Term
| Term | Definition |
|---|---|
| 7 steps to obtaining informed consent | 1.purpose of the study 2.What they will be asked to do 3.anticipated effects 4.How much time it will take 5.What will be done with the data 6.right to refuse to participate 7.can discontinue participation at any time |
| Deductive reasoning | By following downward from the general to the specific, knowledge can be gained about a particular relationship. |
| Inductive reasoning | From specific (accepted knowledge) to general (new knowledge). •Allows new major premises to be determined. |
| Clinical validity | Highest generality to attain in research. Shows that findings can improve quality of life. answers: -Does the effect generalize to a natural setting? -Is it maintained over time? -Does it cause change in person’s social, academic,personal,work life? |
| What is a hypothesis | -Statement of cause-effect relation -Expressed in operational (measurable) terms -Differs from research question by offering a tentative answer -Should be falsifiable. |
| Working hypothesis | Helps guide investigation. “hunch” |
| Research hypothesis | assertion that is seen as highly probable in the light of available evidence and established theory. “educated guess” |
| Null hypothesis | Based on statistical probability. Stated in terms of no relationship between variables or no difference between groups. |
| Alternative Hypothesis | Statement of the expected result of a research study |
| Randomization | The assignment of subjects to experimental and control groups on a random basis with equal probability of being assigned to either group. |
| Simple random Sampling | All subjects in population have an equal probability of being selected. |
| Stratified random sampling | 1st divided into subgroups and then subjects are randomly drawn. (i.e. splitting up into men and women 1st and then randomly choose 10 from each group). |
| Cluster sampling | All subjects are members of a group that was selected randomly (i.e. choosing a school from a school district). |
| Multistage sampling | Select a group and keep choosing to smaller sample (randomly choose a school district, then a school, then a grade than a classroom etc…). |
| Counterbalancing | Experimenter controls all possible sequences of treatments and then randomly assigns subjects to each sequence. More appropriate for between subject design. Not appropriate for within subject designs due to carry over. |
| Single group research | Most common is a Pretest-post test design Often no more than exploratory: –Participants are compared to themselves –However, many uncontrolled variables: •History, maturation; multiple-test effects; statistical regression effects. |
| Between Subject design or two group | -performance of different groups are measured under different treatment conditions and comparisons are made between the groups. Has an experimental and a control group. -Observations made at the same time, so effects are controlled. |
| Within-subject design | Performance of same subjects is compared under different conditions. |
| single-subject | -Focuses on behavior of an individual. -Can be more than 1 subject, but data evaluated individually. -Compare diff conditions(treatment phases) within 1 participant -external validity: replication w/ others -No generalization to pop |
| case studies | -Intensive observation of a person -Often exploratory, but may form basis for further research -No generalization -Multiple case studies may help to form a more general picture -Allows clinicians to compare specific cases to their individual clients |
| Draw backs of case studies: | •Typically have uncontrolled sources of variation •No control group •Inadequate descriptions of dependent and independent variables •Often impossible to replicate •They are not experiments |
| Group design problems | -Don't always lend to treatment efficacy -Samples are usually not random -Sample sizes typically very small -distributions frequently not normal -thus, mean is no longer the sampleâs best representative measure of central tendency repeated measures |
| Mean | Interval or ratio-level measurement that calculates the average score (add scores up and divide by the number of scores) |
| Median | Middle score. -Sometimes preferred if interval or ratio-data are not normally distributed -Rank data (e.g. low-high) to find the ‘middle’ |
| Mode | most frequent- Used only with nominal data |
| Variance | Average variation of scores about the mean. Formula: same as standard deviation but no square root |
| standard deviation | Degree to which scores deviate above or below the mean |
| Range | Difference between the highest and the lowest score |
| Z-scores | Raw scores that have been converted into standard deviation units. •How many standard deviations is this score away from the mean? •They allow us to compare scores across different measurement scales. |
| Percentiles | Z score of 0 means you are at the 50th percentile -take chart score and add .50 to get percentile |
| standard error of the mean | -an estimate of the standard deviation of a sample of means. -allows us to estimate a range of scores within which the true population mean lies. –calculated by dividing the standard deviation by the square root of the sample size |
| confidence interval | confidence about the location of the real population mean. -calculated by subtracting from and adding the standard error of mean to the sample mean |
| Effect size | -An index (or measure) of the degree to which the differences you observed in your sample groups actually exist in the real population. -Greater the ES greater ability to detect a real difference. -As SD incr this deacr = meanB - meanA/SD |
| In a Normal distribution. What percentage of individual's scores fall within one standard deviation of the mean | 68% |
| In a Normal distribution. What percentage of individual's scores fall within two standard deviations of the mean | 96% |
| In a Normal distribution. What percentage of individual's scores fall below the mean and above the mean | 50% |
| In a Normal distribution. What percentage of individual's scores fall more than one standard deviation below/above the mean | 16% |
| In a Normal distribution. What percentage of individual's scores fall more than two standard deviations below/above the mean | 2% |
| Normal Frequency Distributions | mean in the middle of the bell |
| Negatively Skewed | most scores (tail) to left, hump to right |
| Positively Skewed | Hump to left and tail (most scores) to right |
| Bimodal | mean in middle of two humps |
| Nominal Data | –Categorical data, or labels –For example: •cars: Chevy, Ford, Buick •major: Accounting, Biology, Art, History Usually consist of “count data”: –# of cars sold by brand –# of biology majors vs. # of art majors |
| Ordinal Data | -data organized in some hierarchical order -For example: cars arranged by size: VW, Acura, Lincoln -voice rating scale: 1-2-3-4-5 |
| Ratio | data organized in order, with equal intervals between data and an “absolute zero” For example: speed (in miles per hour) score (in number correct) as long as it is possible to get a score of zero …. |
| Independent Variables | –those factors that cause a certain result (factors we usualy try to minipulate or observe) –the things you do to observe an effect (i.e. The treatment we are giving, the age of the group) •the antecedent (what came before) |
| Dependent Variable | –the resulting effect –what change occurred because of what you did –the actual data you are collecting (i.e. fluency rating, improvement on a score on a language path) •the criterion measure (i.e., the consequence) |
| Sensitivity | Kids that are judged to be disordered on our test are actually disordered. |
| Specificity | Normal on our test is also normal based on external measures |
| Equal Appearing Intervals | -Raters make judgments about some condition or situation on (usually) a 3, 5, or 7-point scale. -Scale’s range may be numbers or qualitative labels (descriptors) -It is assumed that the rating scale is an interval measure, & calculations may be made |
| Problems & soultions with Equal Appearing Intervals | -The listeners/raters may not use the entire range of the scale e.g.,End Effect, The listeners/raters may be inconsistent (i.e.poor reliability) Solution: can calibrate listeners judgements by giving examples of different ratings |
| Direct Magnitude Estimation | Listeners are NOT given a scale but use their own unique scale to compile numerical ratings. |
| Paired Comparisons | Rather than assigning each condition a numeric rating, all conditions are presented in all possible pairs. The rater/listener’s job is to determine which member of the pair is better (or worse). |
| Non-Parametric Statistics | -Typically use either the group mode or median as the primary measure of central tendency. *Requirements (assumptions): – Any type of data – Any type of distribution – Equal variances not required |
| Parametric statistics | -Use group mean as primary measure of central tendency -Use if sample size is 11 subjects or more *Requirements: –At least interval data –Data normally distributed –Groups have equal variances -Have greater generalization power to full population |
| Chi2 test | for Cochran Q Test requires a comparison between the actual observed frequency and the expected frequencies χ2 = ∑ (O – E)2/ E |
| Mann-Whitney U-test | –Also known as Wilcoxon rank sum test) –Nonparametric test that tests for differences between groups –Not through their means, but by ranking the obtained scores |
| Interval data | -Data organized in order, with equal intervals between data -For example: temperature (in degrees Fahrenheit or Celsius) or IQ |
| How to calculate percentiles | 1)Locate # trying to get percentile for. 2)If two of same # put 1 in upper group & 1 below. If only 1 # count it as half of value for each group. 3) Count number in the BELOW group 4) Divide this value by the total # of scores 5)x100 for percentile |
Created by:
aramos139
Popular Speech Therapy sets