research design an statistics
Terms
undefined, object
copy deck
- Null hypothesis
- implies that there is no relationship between the independant and dependant variables. Any effects are due to sampling error.
- Alternative hypothesis
- implies tha there is a relationships between the independant and depednent variables
- alternative hypothesis can be either _________or ______
- non directional or directional
- Non directional
- Just states that the null hypothesis is false
- one directional (one tailed)
- not only states that null hypothesis is false but also predicts whether the population parameter estimated by the obtainedsample statistic will be greater or less than the parameter speficied in null hypothesis
- rejection region
- lies in one or both tails of samping distruibution and contains the sample values that re UNLIKELY to occur simply as a result of a sampling eorr. If statistical test indicate that the sample statistic is in the rejection regoin of the sampling distriubution, the null hypothesis is rejected and the alternative hypothesis is retained.
- Retention region
- lies in the central portion of the sapling distrubution and consists of the values that are more likely to occur. . If the sample statistic lies in the rention region, the null hypothesis is retained and the alternative hypothesis is rejected.
- Alpha (level of significance)
- size of rejection region is defined by alpha. if alopha is .05 then five percent of the sampling distribution represents the rejection region and the 985 percent represents the retention region. it is set by the investigator prior to collecting and or analyszing the data.
- statistically significant
- when results of inferential statistical test indicate that the sample statistic lies in the rejection region of the sampling distribution.
- type one error
- occurs when an investigator rejects a true null hypothesis . probability of making at type one error is set at alpha
- Type II error
- occurs when an investigator retains a false null hypothesis
- correct decisions
- retain null hypothesis when it is true or rejectin null hypothesis when it is false
- Power
- when the decision is made to reject a false hypothesis
- parametric test
-
used to evaluate hypothesis about population means variances or other parameters. appropriate when the variable of interest are measured on the interval or ratio scale and when certain assumptions about the popuation are met
Interval ratio data all t-test
one way anova
factoraila anova - Assumption one
- the value of interst is normally distributed in the population
- assumption 2.
- when a study includes more than one group, the cariances of the populations that the different groups represent are assumed to be equal (homoscedasity, refers to equal variances).
- non parametric tests
-
used to analzyze data on variables that are measured on nominal or ordinal scale. Evaluates hypothesis about the shape of the entire distribution rather than the distribution mean's variance or other parameter. Less powerful. ARe less likely to reject a null hypothesis with a non parametric test. Parametric test are preferred
Nominal data: Single sample chi square, Multible sample chi square
Ordinal data: wilcoxon matched pairs test , mann whitney u test
kruskal wallis test. - Critical value.
- used to compare the test statistic that is resulted from a inferentail statistical test (non par or par). if the test equals or exceeds the critical value, the investigator concludes that the obtained sample value lies in the rejection region and the samping distribution and rejects the null hypothesis. if the test statistic is less than the critical value then the rsearcher concludes the sample value lies in the retention region and retains the null hypothesis
- degrees of freedom
- are niumber of values or categories that are free ot vary given the certain values or categories in the distributions are known or fixed.
- Magitude of critical value depends on what two factors
- level of significance and degrees of freedom
- single sample chi test
- when a descriptive study includes one variable and the data to be analyzed are the number of observation in each category: one variable nominal frequency data
- multipl,e sample chi square test
- used when descriptive or expereimental study includes two or more variables and the data to be abalyzed are the number of observations in each category: two or more variables: nominal data
- student's t- test
- used to evaluate hypothesis about the differences between two popuation means as estimated by group means. used to analyze data collected from more than two groups.
- Student's t-test for single sample:
- used when a study includes only one group and the mean obtained by that group will be compared to a population mean. one indepedendant variable single group. One dependant variale, interval or ratio data
- student's t-test for independant sample
- is appropraite statistical test when a study includes two independant groups. one independant variable, two independent groups. one dependent variable interval or ratio data
- Student's t-test for correlated samples
- t- test for correlated samples is used when the two means to be compared have come from correlated groups.for exaple a study using within subject design in which a single group of subjects will be compared to itself. or when subjects are matched on extraneous variable and members of each matched pair have been assigned to different groups.
- Anova
- is used to compare two or more more means. It tests hypothesis that the populatio means estimated by the group means are equal.
- One way anova.
- appropriate statistical test when a study includes one independant variable and two or more independent groups. LIke the t0test, the anova compres means but does more complex way by analyszing variablity around the means. one indepedent variable, two or more indepednt groups one dependent variable, interval or ratio data
- Factorial, analysis of variance
- appropriate statistical test when a stuy includes two or more independant variables. Two ivs: two way anova. two or more independent variables , independent groups .One dependnt variable interval or ratio data.
- regression analysis
- assumption underlying regression analysis is that there is a linear relationship between the two varaibles of interst and therefore, that the relationship can be sumarized by a straight line: the relation ship can be sumarized by a straight line, and the relationship can be described by a regression line or line of best fit.
- longitudinal studies
- investigate changes by assessing the same group of people over an extended period of time. Overcome the problem of cohort effects but difficult to conduct because of time and cost factors and subject to high drop out rates
- Cross sequential studies
- combines cross-secttional studies and longitudinal methodologies by asessing member of two or more age groups at two or more different times.
- Evaluation research
- conducted to assess the value or effectiveness of psychological educational and other social programs.
- Extraneous variable
- irrelevant to the purpose of a research but counfounds the study's results because it has a systematic effect on the dependant variable
- Control extraneous variables
-
1.random assignment of subjects to a treatment group
2.hold extraneous varaible constant (selecting subjects who are homeogenous wiht the respect to that variable)
3. Match subjects on teh extraneous variable (match subjects with that variale then randomly assign matched subjects to one of the treatment groups.)
4. building extraneous variable into the study (include as independant variable)
5. Satistical control of the extraneous variable
(ancova) or other statistical tecniques to remove varaiblity in the dependant variable. - Threats to internal validity
-
1.maturation
2. history
3.testing
4.instrumentation
5. statistical regression
6.selection
7. attrition (mortality - Threats to external validity
-
1. interaction between testing and treatment. When administration of a pretest can sensitize subjects
2. interaction between selection and treatment. (volunteers are different from non volunteers.)
3. Reactivity: Reacting a particular way just because you know you're being observed . Hawthorne effect, evaluation apprehension, demand characteristics
4.multiple treatment interference. Occurs when exposed to several levels of treatment. - Hawthorne effect
- tendenct of subjects to perform better because the attention they are recieving as research partcipants
- Evaluation apprehension
- causes subjects to act in ways they believe will help them avoid negative evaluation
- Demand characteristics
- behavior or research participants can be altered by cues that inform the subjects of the purpuse of the study or suggest what behaviors are expected of them
- Experimenters expectancy
- Can unintentionally provide subjects with cues
- Between group designs
- Administer different levels of independant variable are administered to diffent groups of subjects and then comparing the status or performance of the groups on the dependant variable.
- within subjects designs
-
all levels of the independant variable are administered at differnt times to all subjects
comparisons are made within subjects rather than between groups of subjects - Mixed designs
- combines between groups and wihtin subjects methodology
- counterbalanced deisngs
- mixed desings because they permit comparisons both between groups and within subjects
- Single subject designs
- Inclue at least one base line phase and one treatment phase. subjects act as their own no treatment controls
- AB design
- single baseline and single treatment phase
- ABA, ABAB design
- requires withdrawl of the treatment during second and subsequent baseline phases. Reversal or withdrawl design
- multiple baseline desgn
-
does not require the withdrawl of a treatment during the course of the study . involves sequentially applying the treatment to different behaviors of the same subject, to the same subject in differtn setting
to the same behavior in differnt subjects . - Nominal scale
-
divides varibales into uordered categories
sex of people. eye color, polititical affliation - Ordinal scale
- divides observations into categories and provides information on the order of the category. Use of rating scales
- Interval scale
- property of order as well as the property of equal intervals betwen successive points on the measurement scale
- Ratio scale
- properties of order , equal intervals and property of an absolute zero point. Score of zero indicates a complete absence
- Kurtosis
- relative peakness of a distribution
- Lptokurtic
- distribution is more peaked then the normal curve
- Platykurtic
- Distrubution is flatter than the normal curve
- Mesokurtic
- normal cuve
- Postively skewed distribution
- most of the scores are in the negative side of the distribution. Positive tail is extended because of a presense of a few high scores
- Negatively skewed distribution
- most scores are on positive side with the tail being in the negative
- Mode
- score or category that occurs most frequency in a set of data
- median
- score that divides the distribution of data in half when data have been ordered from low to high. Equal to the middle observation
- Arithmetic mean
- average .summing all values and dviding the sum of of values in the distibution
- Choosing a measure of central tendency for nominal, ordinal ,interval and ratio
-
Nominal: mode
Ordinal: mode or median
Interval and ratio: mode, median or mean - Meausres of variablity
-
1. range
2. Variance
3. standard deviation - Range
- caluculated by subtracitng the lowest score in the distribution from the higher score. can be misleading if there is a really high score or really low score that is not typical
- Variable
- provides measure of average amount of varaiblity in the distribution. Includes all the scorews in distribution. variance requires squaring each deviation score . used in inferential statistics
- Standard deviation
- expressed in the same unit of measurement as original scores. Calculated by taking the square root of the variance
- mann whitney U test
- study inclues two independant groups and data on dependant variable are reported in terms of ranks. assesses whether or not the ranks of the observation in one group are equivielant to the ranks of obervations in the other group.
- Wilcoxon matched pairs signed ranks test
- when a study includes two correlated groups and the differences between the dependant variable scores are converted to ranks . assess whether or not there is a difference bewteen the sum of the ranks of the positive difference between socores and the sum of the ranks of the negative diffences between them
- Kruskal wallis one way analysis of variance test
- similar to man -whitney u test but used when a study includes two or more independnt groups . Test is used to assess the hypothesis that the ranks of observations in the differnt groups are equivant.