Research Flashcards
What are the two types of experimental design research?
Experimental research with randomization (True experimental design) (RCT)
Experimental research without randomization (Quasi-experimental)
What is true experimental research?
Manipulation
Randomization
Control of other variables
What is quasi-experimental study?
There is no random assignment. Participants may be assigned based on pre-existing conditions (e.g., by age, location, or school), which increases the likelihood of biases
What are the non-experimental (observational) studies?
Cohort (has the highest level of evidence among observational studies)
Case-control (moderate evidence)
Cross-sectional (least level of evidence)
What is cohort study?
Longitudinal time (mahabang gagawin)
In a cohort study, researchers follow a group of people (a cohort) over time to see how different exposures (e.g., smoking, diet) affect outcomes (e.g., disease development). The cohort is usually divided into exposed and non-exposed groups.
What are the two types of cohort study? What is the difference between the two?
Prospective cohort: starts in the present and follows participants into the future, collecting data over time → forward in time
Retrospective: looks back at data that already exists and examines participants’ past exposures and outcomes ← backward in time
What is a case control study?
Compares subjects with disease to those without
A case-control study is a type of observational research used to investigate the causes of a specific outcome (such as a disease or condition) by comparing two groups of people: those who have the outcome (cases) and those who do not (controls).
Most of the time we use this study for rare diseases
What is cross-sectional study?
Presence/absence of disease at a single point in time
Example for this study: prevalence
Give an example for a cohort study
Incidence (rate of new cases) of childhood apraxia of speech in preschool-aged children over a one year period
Effectiveness of early language intervention
What is the baseline?
Refers to the initial set of data collected before any interventions, treatments, or experimental changes are applied. It serves as a reference point or starting condition that allows researchers to compare the “before” and “after” effects of an intervention or change in an experiment.
Give an example of a case-control study
A study on risk factors for childhood stuttering
Give an example of cross-sectional stud
Prevalence (number of cases with the disease) of voice disorders among public school teachers
What is quantitative and qualitative study?
Quantitative study answers questions through standardized measures
What is qualitative study?
Answers questions through in-depth interviews, focused group discussions, and observations
What is reliability?
Consistency, repeatability
Test-retest → over time (consistent results over time)
Internal consistency → within a measure/instrument (there is one measurement; the questions or sub questions are consistent)
Inter-rater → across different raters/observers (consistent and judgment even if there are different raters)
Intra-rater → same rater (same observer but has consistent results)
What is validity?
Accuracy
Face → what is superficially appears to measure (based on looks e.g., weight scale)
Content → covers all aspects (Example: A math test intended to measure overall math ability should include questions covering all key areas of math (e.g., algebra, geometry, arithmetic). If it only tests algebra, it lacks content validity.)
Construct → existing theory of knowledge (is it consistent with the theory behind the concept) ( Example: construct validity is like making sure your test really measures what it’s supposed to. Just like with the drawing test, you want to check that you’re asking the right questions and that the results make sense based on what we already know about drawing.)
Criterion → gold standard of measure (This measures how well one test or measurement compares to a gold standard or established benchmark for the same concept. It assesses the test’s performance against a criterion that is already known to be valid.)
What is the validity of diagnostic tools?
Sensitivity
Specificity
What is sensitivity?
Indicated how well the tool identifies a disease
What is specificity?
Indicates how well the tool identifies absence of a disease
What is a positive test result (+)?
With disease (true positive → sensitive)
Example:
Scenario: Imagine a group of 100 people, and 10 of them actually have COVID-19 while the rest do not.
High Sensitivity: If the PCR test has a sensitivity of 90%, this means that it correctly identifies 9 out of the 10 people who have the virus.
Positive Test Result (Sensitivity): So, if one of those 10 infected individuals takes the test, there is a high likelihood (90%) that the test will come back positive for them, indicating that they have COVID-19.
What is a negative test result (-) or true negative (specificity)?
Example:
Scenario: Imagine a group of 100 people, and 10 of them actually have COVID-19 while the rest do not.
High Sensitivity: If the PCR test has a sensitivity of 90%, this means that it correctly identifies 9 out of the 10 people who have the virus.
Positive Test Result (Sensitivity): So, if one of those 10 infected individuals takes the test, there is a high likelihood (90%) that the test will come back positive for them, indicating that they have COVID-19.
Negative Test Result/True negative (Specificity): The remaining 90 people (who do not have COVID-10) all test negative
What is probability sampling?
Randomization
Simple → each member has an equal chance of being selected (Example: If there are 100 students in a school, each student has a 1 in 100 chance of being selected for a survey.)
Systematic → selecting every nth member from a population
Stratified → dividing the population into homogenous groups (strata) then randomly selecting samples from each group (Example: If a researcher wants to ensure they have an equal representation of different age groups in a study, they might divide the population into age brackets (0-18, 19-35, 36-50, etc.) and randomly select participants from each age group.)
Cluster → dividing the population into clusters (usually geographical) then selecting some clusters (Example: A researcher studying student performance might divide a city into districts (clusters) and randomly select a few districts to survey all the students within those districts.)
What is non-probability?
No randomization (Quasi experimental design)
Convenience → based on researcher’s easy accessibility (Example: A researcher might survey people in a nearby coffee shop because it’s convenient for them to reach those individuals, rather than seeking a more representative sample from the entire population.) what’s convenient for the researcher
Purposive → based on the researcher’s judgment (Example: A researcher studying a specific medical condition may select individuals who have that condition rather than randomly sampling the general population, believing that these individuals will provide more relevant insights.)
Quota → predetermined number (Example: If a researcher wants to include 50 men and 50 women in a study, they will continue to sample until they have 50 participants from each gender, regardless of how they choose those individuals.)
Snowball → word of mouth; referrals from existing participants (Example: A researcher studying a rare condition might initially interview one or two individuals who have it. Those participants then refer others they know with the same condition, helping the researcher expand their sample through these connections.)
What are tables?
Present data in rows and columns
Compare values
Purpose:
They allow for easy comparison of values across different categories or groups.
Tables can summarize large amounts of data in a clear and concise manner, making it easier for readers to find specific information.
Example: A table might show the test scores of students across different subjects, where each row represents a student, and each column represents a subject.
What is a bar graph?
Represent categorical data with rectangular bars of varying lengths
Purpose:
Bar graphs are useful for comparing different categories or groups.
They visually highlight differences between categories, making it easy to see which category has the highest or lowest values.
Example: A bar graph could show the number of students enrolled in different sports at a school, with each sport represented by a bar.
What are pie charts?
Represent the proportion of each category within a dataset
Purpose:
They visually show the relative sizes of parts to a whole.
Pie charts are effective for displaying percentages or proportions and making it easy to see how different categories contribute to the overall total.
Example: A pie chart might show the market share of different smartphone brands, with each slice representing the percentage of total sales for each brand.
What is a histogram?
Useful for understanding the shape and spread of data distributions
Purpose:
Histograms are useful for understanding the shape and spread of data distributions (such as normal distribution, skewed distribution, etc.).
They can reveal patterns, such as where data points cluster or if there are any outliers.
Example: A histogram might show the distribution of test scores in a class, with bins representing score ranges (e.g., 0-10, 11-20, etc.).
What is a scatter plot?
Useful for visualizing relationship between continuous variables
Purpose:
Scatter plots are useful for visualizing the relationship between two variables, such as whether they correlate positively, negatively, or not at all.
They can also help identify trends, clusters, or outliers in data.
Example: A scatter plot might show the relationship between hours studied and exam scores, with each dot representing a student’s hours studied (x-axis) and their corresponding exam score (y-axis).
What are descriptive statistics?
Summaries and described the characteristics of a dataset
Methods: measures of central tendency, measures of variability, skewness
What is inferential statistics?
Makes inferences or predictions about a population based on data
Methods: hypothesis testing, regression analysis
Descriptive Statistics → Measures of central tendency
Measure of central tendency indicate where the center of a dataset lies
Mean - Average
Median - middle value
Mode - value that appears most frequently
Descriptive Statistics → Measures of variability aka Measures of dispersion
Measures of variability describe how spread out the data is
Range - difference between highest and lowest value in dataset
Variance - the sum of squared deviation from the mean
Standard deviation - the square root of the variance, indicating the average distance of each data point from the mean.
Descriptive statistics → skewness
Skewness indicates the asymmetry of the data solution
Zero/no skew (normal) also referred to as no skew or normal distribution) indicates that a dataset is perfectly symmetric around its mean; bell-curve–A smooth, bell-shaped curve centered around the mean (the mean
Positive skewness indicates a longer tail on the right side (more low values) mean is higher, mode is the lowest
Negative skewness indicates a longer tail on the left side (more high values) mode is the highest, mean is the lowest
The higher the variance of a dataset, the more ________________
The higher the variance of a dataset, the more spread out the data points are.
nferential testing → hypothesis testing
Null hypothesis - Represents no effect (the null hypothesis is a statement that indicates no effect, no difference, or no relationship between variables. It serves as the default position that there is nothing new happening in the population.
Example of null hypothesis: “There is no difference in test scores between students who study with a tutor and those who do not.”
Alternative Hypothesis (H1 or Ha) - Represents what the researcher aims to prove or support. It suggests that there is an effect, a difference, or a relationship.
Example of alternative hypothesis: “Students who study with a tutor have higher test scores than those who do not.”
What is the significance level (alpha)
You select the significance level after formulating the hypothesis (whether it is a null hypothesis or an alternative hypothesis)
0.05 - There is a 5% chance of rejecting null hypothesis when it’s true
0.01 - There’s a 1% chance of rejecting null hypothesis when it’s true
What is regression analysis?
Analyzes relationship between variables (a dependent and one or more independent variable)
What are the research variables?
Dependent - outcome, response, effect (think of this as the result of the study; Example: (1) In a study on how different amounts of sleep affect test scores, the dependent variable is the test score. (2) In an experiment to see if a new medication reduces pain, the dependent variable is the level of pain reported by patients.)
Independent - treatment, manipulated, controlled, cause/predictor (Think of it as the “cause” or “predictor” in an experiment. Example: In the sleep study, the independent variable is the amount of sleep each participant gets (e.g., 4 hours, 6 hours, 8 hours).
Control - aka confounding variables, constant (these are variables that are kept constant throughout the study to prevent them from affecting the results (dependent variable). These are often called confounding variables if they are not controlled and end up influencing the results unintentionally. Example: In the sleep study, a control variable could be the type of test used, since you want to ensure that any differences in test scores are due to sleep and not different types of tests; In the medication study, a control variable might be age or diet, as you want to make sure that age or diet isn’t affecting the pain levels, only the medication.)
State the independent, dependent, control variables: A researcher wants to find out whether increasing daily exercise leads to more weight loss in adults. They ask participants to engage in different amounts of exercise per day (30 minutes, 1 hour, and 2 hours) and measure their weight loss after a month.
Dependent: result of the weight/ weight loss
Independent: amount of daily exercise (30 minutes, 1 hour, 2 hours)
Control variable: Diet, age, and initial weight of participants (kept constant)
State the independent, dependent, control variables: A psychologist investigates whether the number of hours of sleep impacts students’ performance on math tests. They divide students into groups who sleep 4 hours, 6 hours, and 8 hours per night for a week, then give them all the same math test.
Independent Variable: Number of hours of sleep per night (4 hours, 6 hours, 8 hours)
Dependent Variable: Scores on the math test
Control Variables: The difficulty of the math test, study time, and prior math ability of students (kept constant)
State the independent, dependent, control variables: The effect of articulation therapy techniques on speech intelligibility in adults with dysarthria
Dependent: Speech intelligibility in adults with dysarthria
Independent: Articulation therapy techniques
Constant variable: Age, severity of dysarthria
What is p-value?
The p-value indicates the probability of obtaining the observed results (or more extreme) given that the null hypothesis is true (it is a measure in statistical hypothesis testing that helps determine the significance of the results obtained from a study or experiment–basically, it helps determine whether the result of a study are statistically significant)
A smaller p-value indicates strong evidence against the null (supporting the alternative hypothesis → rejecting the null)
A larger p-value suggests weak evidence against the null (accept the null)
So if the p-value is below or equal the significance level (alpha), reject the null hypothesis. Example: If the p-value is 0.03 and α is 0.05, you would reject the null hypothesis because the p-value is less than 0.05, meaning the results are statistically significant → meaning the evidence is strong enough to support the alternative hypothesis
However, if the p-value is more than the significance level (alpha), it fails to reject the null hypothesis meaning there is not enough evidence to support the alternative hypothesis or in simpler terms, you have to accept the null; Example: If the p-value is 0.08, you would fail to reject the null hypothesis since the p-value is greater than 0.05, meaning the results are not statistically significant (accept the null because the evidence is too weak)
Try to solve this: Suppose a researcher is testing whether a new drug lowers blood pressure compared to a placebo. The null hypothesis (H₀) is that the drug has no effect on blood pressure. After collecting data, they calculate a p-value of 0.02. If the significance level (α) is set at 0.05, what does this mean?
Since the p-value is lower than the significance level, the clinician or researcher will reject the null hypothesis (meaning the drug has an effect on the person–there is an effect; the evidence is strong enough against the null → high evidence to support the alternative hypothesis)
Try to Solve this: Null hypothesis - Peppermint essential oil has no effect on the pangs of anxiety. Alternative hypothesis - Peppermint essential oil alleviates the pangs of anxiety. (Significance level is 0.05, p- value that the researcher has is 0.06)
Since the p-value is higher than the significance level, the researcher must accept the null hypothesis (rejecting the alternative hypothesis)
There is weak evidence that the peppermint essential oil has an effect.
There is strong evidence that the peppermint has no effect → accepting the null
What is type 1 error?
False positive → the null was rejected but it is actually true (the null should have been accepted but instead it was rejected)
What is type 2 error?
False negative → the null hypothesis is accepted when in fact it should be rejected; missed opportunity “akala mo walang effect pero MERON MERON”
State whether it is a type 1 or type 2 error: Hypothesis: SLPs are poor in mathematics. Conclusion: Hypothesis is correct and that SLPs are poor in mathematics when in reality you should have rejected your hypothesis.
TYPE 2 ERROR
EBP is an integration of
Expertise of clinician
Best evidence available
Patient’s preference
What are the components of EBP?
Client perspectives
Clinical expertise
Evidence (External and Internal)
What is the hierarchy of evidence (from highest level of evidence to lowest)?
(1) Systematic Reviews & Meta-analyses
(2) Randomized Control Trials (RCTs)
Observational Studies:
(3) Cohort Studies
(4) Case-control studies
(5) Cross-sectional studies
(6) Case Reports
What is a systematic review?
It is a type of research study that synthesizes/summarizes and evaluates all available evidence on a particular research question or topic. It follows a structured and transparent process to ensure that the review is comprehensive, unbiased, and repeatable.
What is meta-analysis?
It is a statistical technique used to combine and analyze data from multiple studies on the same topic. It is often done as part of a systematic review and helps to produce a more accurate and comprehensive answer to a research question by increasing the sample size and overall power of the analysis.
→ A meta-analysis takes the results of several independent studies (usually quantitative) and integrates them into a single analysis. This allows researchers to pool data from different sources, even if individual studies had smaller sample sizes or varied slightly in their methodology
What are the appropriate research designs for common types of research questions?
Descriptive research questions → Observational studies
Experimental research questions → RCTs, Quasi-experimental design
Identify associations → correlational study
Qualitative Research Questions → qualitative research methods such as interviews, focus groups, ethnography
What are the steps for the EBP Process?
(0) EBP Basics
(1) Ask - ASK: Clinical question guideline: PICO/PICOTT (Population, Intervention, Comparison, Outcome, Time, Type of Study)
(2) Acquire - Evidence: Search!
(3) Appraise - Validity and relevance
(4) Apply - Patient care decisions
(5) Assess - Performance evaluation, patient outcome review
Common avenues
Conferences and symposia
Scholarly journals
Professional associations
Books
Theses
Government agencies, policymakers, non-profit organization
Community organizations
Online platforms/social media
What are the steps of publishing research?
Selection → choose a journal
Preparation → write your research paper
Submission → Submit your paper
Peer-review process → Your paper will undergo peer-review after your submission
Revision → Address any feedback or revisions requested by reviewers or editors
Acceptance/rejection → the journal system will inform you whether your paper has been accepted or rejected
Proofreading and finalizing → once accepted for publication, make any necessary final revisions
Publication → research paper will be published after final approval
Promoting and sharing → increase the visibility and impact of paper by sharing it on relevant platforms
Post-publication communication → Engagement with readers
Who are the members of the research team?
Principal investigator → leads the research project and oversees the team
Co-investigator → assist the principal investigator
Research assistants → Support the research project by performing tasks like data collection, and administrative duties
Statisticians → Provide statistical expertise
Subject matter experts → Contribute expertise in a particular field relevant to the research project
Collaborators → External individuals like peer-reviewers
Ethics Committee Members →
Ensure that the research complies with ethical guidelines and regulations
Funding Agencies → Provide financial support
Project Managers → Coordinate and manage day-to-day activities of the research project (timeline, budget, etc.)
Research participants → Participants of the study
The null hypothesis was rejected when it should have been accepted. What type of error is this?
Type I error: False positive
The null hypothesis is accepted when it should have been rejected.
Type II error. False negative
An SLP has wrongfully interpreted data collected during a recent study and has claimed that the treatment leads to participant progress when no progress was actually made. What type of error is this?
Type I error: False positive
The clinician hypothesized that there is no effect in the intervention when there was actually progress because of the intervention. What type of error is this?
Type II error: False negative
A researcher is conducting a clinical trial comparing two interventions for speech therapy in children with language delays. To minimize bias, the researcher decides to randomly assign participants to either treatment group. What type of research design is being used in this study?
A. Cohort study
B. Randomized controlled trial (RCT)
C. Case-control study
D. Cross-sectional study
Answer: B. Randomized controlled trial (RCT)
Rationale: In an RCT, participants are randomly assigned to different intervention groups, which helps control for confounding variables and bias, making it the gold standard for comparing interventions.
A clinician wants to investigate the impact of early language intervention on vocabulary development. The study includes a pre-test before intervention, a post-test after intervention, and a follow-up test 6 months later. What type of statistical analysis is most appropriate to examine changes over time within the same group?
A. Independent samples t-test
B. Repeated measures ANOVA
C. Chi-square test
D. Pearson correlation
Answer: B. Repeated measures ANOVA–
Rationale: Repeated measures ANOVA is used when the same participants are measured multiple times (pre-test, post-test, follow-up) to analyze changes within a group over time.
In a research study, the null hypothesis states there is no significant difference in language outcomes between two speech therapy techniques. The p-value obtained is 0.03. What can be concluded if the significance level (α) is set at 0.05?
A. The null hypothesis is accepted, and there is no significant difference.
B. The null hypothesis is rejected, and there is a significant difference between the two techniques.
C. The alternative hypothesis is rejected, and the results are due to chance.
D. The null hypothesis cannot be rejected because the p-value is not less than 0.01.
Answer: Answer: B. The null hypothesis is rejected, and there is a significant difference between the two techniques.
Rationale: A p-value of 0.03 is less than the significance level of 0.05, so the null hypothesis is rejected, indicating a statistically significant difference between the speech therapy techniques.
A study is examining the relationship between frequency of therapy sessions and improvement in communication skills. After controlling for age and severity of the disorder, the researcher finds a strong positive correlation. Which of the following statistical methods is best suited for this type of analysis?
A. Logistic regression
B. Simple linear regression
C. Multiple regression analysis
D. Factor analysis
Answer: C. Multiple regression analysis
Rationale: Multiple regression is appropriate for assessing the relationship between a dependent variable (improvement in communication skills) and multiple independent variables (frequency of therapy, age, severity of the disorder), controlling for confounding factors.
In an experiment on a new stuttering intervention, participants are split into two groups: one receiving the new intervention and the other receiving standard therapy. After the treatment, the new intervention group shows significant improvement. However, an external auditor points out that participants were not blinded to the intervention. What kind of bias is most likely to have affected the results?
A. Selection bias
B. Recall bias
C. Performance bias
D. Attrition bias
Answer: C. Performance bias
Rationale: Performance bias occurs when participants or researchers are aware of which intervention is being received, potentially influencing the outcomes. In this case, participants not being blinded could have led to biased results.
A researcher uses a quasi-experimental design to assess the impact of a language therapy program in a school setting. However, the school district only allows implementation in schools with higher resources. What is the primary threat to the internal validity of the study?
A. History effect
B. Selection bias
C. Maturation effect
D. Testing effect
Answer: B. Selection bias
Rationale: Selection bias is a threat when groups being compared are not equivalent from the start. In this case, schools with higher resources may differ significantly from other schools, introducing bias in the results.
In a systematic review, researchers aim to evaluate the overall effectiveness of a treatment by combining data from multiple studies. They notice substantial variability in the effect sizes across studies. What statistical method should they use to account for this variability?
A. Fixed-effects meta-analysis
B. Random-effects meta-analysis
C. ANOVA
D. Meta-regression
Answer: B. Random-effects meta-analysis
Rationale: A random-effects meta-analysis accounts for variability (heterogeneity) across studies by assuming that the effect sizes are not identical but vary from study to study, making it appropriate when there’s substantial variability.
A speech-language pathologist wants to investigate whether the severity of aphasia is associated with the success rate of a specific treatment. The researcher collects data from 30 patients and categorizes the success rate as “success” or “failure.” Which statistical test would best analyze the relationship between these categorical variables?
A. Chi-square test
B. Independent t-test
C. Mann-Whitney U test
D. Linear regression
Answer: A. Chi-square test
Rationale: The chi-square test is used to examine the association between two categorical variables (in this case, severity of aphasia and success/failure of treatment).
A researcher finds a statistically significant result in their study of a new intervention for childhood apraxia of speech. However, the effect size is small. Which of the following best describes the clinical significance of the finding?
A. The result is likely to have a large impact in clinical practice.
B. The result is statistically significant but may have limited practical importance.
C. The finding is invalid due to the small effect size.
D. The result should be disregarded because statistical significance does not matter.
Answer: B. The result is statistically significant but may have limited practical importance.
Rationale: A statistically significant result with a small effect size suggests that while the result is unlikely to be due to chance, it may have limited clinical or practical relevance, as the effect is small.
In an observational study examining the prevalence of a communication disorder across different age groups, the researcher collects data from participants at a single point in time. Which research design does this study represent?
A. Longitudinal study
B. Experimental study
C. Cross-sectional study
D. Case series
Answer: C. Cross-sectional study
Rationale: A cross-sectional study collects data from participants at one point in time to examine the prevalence of a condition or association between variables, without following them over time.
