HLTH 240b Lecture Notes - Lecture 10: Analysis Of Variance, Random Assignment, Measuring Instrument

59 views4 pages

ultramarinealligator305

18 May 2017

School

Yale University

Department

Global Health Studies

Course

HLTH 240b

Professor

Marney White

For unlimited access to Class Notes, a Class+ subscription is required.

Document Summary

If the p value is small, then it is unlikely that the differences you observed are due to random sampling. One way anovas: one way because only one independent variable, can have many levels of the independent variable. Ans arousal or personality: female participants, dv= attractiveness ratings, two ivs. Iv-1 (movie type) 2 levels: scary movie vs. documentary. Iv-2 (personality) 2 levels: jerk versus nice guy. Lines cross: suggests interaction, meaning: influence of one variable depends on the level of the other variable, effect of personality depends on the type of movie: if scary movie, women prefer a nice guy. If documentary, there is no effect of personality. Interactions effect trumps main effect: main effect for nice guy or not, but switches depending on type of movie watched. Intervention given: collect data after intervention, after-only non-equivalent control group design, difference is no info on pre-treatment or baseline features. Infers: people may self-select into certain groups, example.

Related Questions

Lab Topic 1: The Process of Science

Introduction

1. Which of the following questions would you consider appropriate scientific questions that can be tested using the scientific method? For the statements that can be tested using the scientific method, explain how your experiment would be set up. Explain the different steps in the scientific method as they relate to this statement.

a. Does the pH of the soil affect the color of hydrangea flowers?

b. What properties of the soil influence the color of hydrangea flowers?

c. Do hydrangeas look more beautiful when grown in the right soil?

d. Was The Day After Tomorrow a better movie than Terminator?

e. Does grass require fertilizer in order to grow?

2. Read this hypothesis and answer the questions below:

Plants kept closer to windows grow larger than plants kept several feet away from windows.

a. Indicate the dependent variable:

b. Indicate the independent variable:

c. List 3 other variables that must be kept constant in this experiment (the control variables).

d. Write a prediction based on the above hypothesis.

Exercise 1.1

Take a Deep Breath

Students will need to take a picture of themselves or their group walking up and down on the step or doing other exercise. Include a note with your name and date on an index card in the picture. Insert picture here:

1. Identify the following in the experimental design. You will need to refer to the protocol to answer this question.

a. Independent variable:

b. Dependent variable:

c. Experimental dataset:

d. Control dataset:

e. Hypothesis:

2. Fill in the data for each individual in your group for the Breath experiment.

	Length of time breath can be held (seconds)
Individual Tested	Resting	After 1 minutes of exercise	After 5 minutes of exercise




Average (take the average for each column)

3. Look at the averages for all individuals in your group. Was the hypothesis you stated in the previous question supported by the âaverage dataâ? Explain your answer!!!

4. Now look at the data for each individual separately. Was there data for anyone in your group that did not support the hypothesis? If yes, explain whythis could be.

5. How many replicates did your group have?________________ Why is it good to have more than one replicate?

Exercise 1.2

Letâs Graph

In this exercise, you will be graphing the data that you collected in exercise 1.1.

1. In exercise 1.1, you identified the independent and dependent variables. Which one goes on the X axis of the graph and which goes on the Y axis?

2. Will you make a line or a bar graph of your data? Explain why you chose this type of graph.

3. You will need to use excel to generate a graph of your data. On this graph, you will need to label the title, independent variable, dependent variable and possibly the legend. You will need to select an appropriate scale so the data is spread out on the graph.

*Consult with your instructor to see how he/she would like for you to submit your graph. You might work on it in class and submit it through D2L or you might submit it at a different time. If you are taking the online version of the class, you will need to use Excel or word to make the graph.

Exercise 1.3

You will need to get the data under exercise 1.3 in the protocol sheet to complete these sections.

1. What is the dependent variable?

2. What is the independent variable?

3. Which, if any, of these individuals has diabetes? ___A or B ___?

4. What data do you have to support your hypothesis?

5. Construct a graph showing the time after eating each group had the greatest blood glucose levels.

Chose the correct answers on one through 15(plese help)

1.A research group hypothesizes that successful infection by HIVdepends upon the expression of FSH3, a human gene. They overexpressthis gene in T cells (Which HIV normally infects) so that FSH3protein levels are very high in these cells. In other T-cells, theymutate FSH3 so that none of its encoded protein is made. They thenmeasure HIV infection rate in the control, overexpression, andmutant T cells. The three cultures are otherwise grown in the sameway.

What is the independent variable(s) in this experiment? (Selectall correct answers)

		FSH3 protein levels
		The medium in which cells are grown
		The temperature at which cells are grown
		HIV resistance of cells

Question 2

Which arises from an error in a measurement or observation?

		Technical variability
		Biological variability
		Standard error
		Standard deviation

Question 3

Use the following description of an experiment to answer thenext question.

A research group hypothesizes that successful infection by HIVdepends upon the expression of FSH3, a human gene. They overexpressthis gene in T cells (Which HIV normally infects) so that FSH3protein levels are very high in these cells. In other T-cells, theymutate FSH3 so that none of its encoded protein is made. They thenmeasure HIV infection rate in the control, overexpression, andmutant T cells. The three cultures are otherwise grown in the sameway.

What is the control group(s) in this experiment? (Select allcorrect answers)

		The unaltered T cells
		The medium in which cells are grown
		The T cells in which FSH3 is mutated
		The T cells overexpressing FSH3

Question 4

How many of the following statements are true?

I. All papers have an abstract, methods, results, and discussionsection

II. The methods always follows the introduction

		0
		2
		1

Question 5

If scientific convention states that when a P value of less than0.05 is considered significant (in other words, the two groupsbeing measured are actually different), this means that: (Selectall correct answers)

		If you were to repeat the experiment 100 times, there would notbe a significant difference between the groups in 5 of theexperiments
		95% of all conclusions that state two groups are not differentbased on a 0.5% p-value threshold are false negatives
		The groups are actually not different from each other 95% of thetime
		5% of all conclusions that state two groups are different basedon a 0.5% p-value threshold are false positives

What information is typically found in the introduction sectionof a paper? (Select all that apply)

		The dependent and independent variables within the study
		Which areas of the field are not well understood and whatquestion the study addresses
		How the main question of the paper is related to other researchin the field.
		How previous work forms the basis for the current study
		The purpose of the study
		Details of the approaches used
		A summary of relevant related studies

Question 7

How can you judge for yourself whether the conclusions reachedin a paper are valid? (Select all that apply)

		Read papers of other authors studying the same question or areaof inquiry
		Determine if each hypothesis is supported by different types ofevidence/approaches.
		Consider closely whether the data actually both addresses andsupports the authorâs hypothesis.
		Examine the amount of data in the paper, and how long the paperis.
		Whether the authors find a mechanism, or exactly how on variableaffects another.
		Interpret the data for yourself without looking the discussion,then compare your interpretations to that of the authorâs.

Question 8

Where can you usually find the institutional affiliations (wherethey work) of the authors?

		Within the discussion
		In the acknowledgements section
		Below the authors list
		This is not usually included in papers

Question 9

Other than publishing primary research articles, what is themain way scientists disseminate their results?

		Lab meetings
		Personal correspondence
		Conference seminars
		Public lectures

Question 10

Use the following description of an experiment to answer thenext question.

What is the experimental group(s) in this experiment? (Selectall correct answers)

		Each group is its own control
		The T cells overexpressing FSH3
		The T cells in which FSH3 is mutated
		The unaltered T cells

grayclam130

Read the article below and write a brief (1-2 paragraph) summary. This assignment should be approximately 250 words . Complete this assignment as a Word document.

Time to Scrap Performance Appraisals

Times have changed. More and more companies have decided to radically change (read âscrapâ or âre-engineerâ) their performance appraisal process.

Last week at our research conference we spoke with Adobe, Juniper, Kelly Services, and a variety of other companies who have decided to do away with traditional performance ratings and dramatically change the annual appraisal process.

Our research shows that this is a strong and positive trend.

Why the process must change.

Why do companies have annual reviews in the first place? They are an artifact from traditional top-down organizations where we had to âweed outâ the bottom performers every year. By forcing managers to rate people once per year we can have annual talent reviews and decide who gets more money, who to promote, and who to let go.

Coupled with the performance rating is the âpotentialâ rating, which tries to capture an individualâs potential to move up two levels in the organization (the traditional definition).

This approach is based on a philosophy that â we cant totally trust managersâ so weâre going to force them to fit people into these rating scales. And in many companies (around 20%) there are forced distributions, which mandate that some percent of employees are rated at the bottom and only a limited percent can be rated at the top.

The well publicized problems with this process abound. These include:

Employees need and want regular feedback (daily, weekly), so a once-a-year review is not only too late but itâs often a surprise. Regular coaching is the key to alignment and performance.

Managers cannot typically âjudgeâ an entire year of work from an individual at one time (imagine if your spouse gave you an annual review!), so the annual review is awkward and uncomfortable for both manager and employee.

Manager-employee relationships are not 1:1 like they used to be. We work with many leaders and peers during the year, so one person cannot adequately rate you without lots of peer input.

While some employees are a poor fit and likely are poor performers, these issues should be addressed immediately, not at the end of the year.

Some companies really do have a lot of high performers, so forced ranking eliminates great people and damages the culture.

People are inspired and motivated by positive, constructive feedback â and the âappraisalâ process almost always works against this.

The most valuable part of an appraisal is the âdevelopment planningâ conversation â what can one do to improve performance and engagement â and this is often left to a small box on the review form.

Companies are nervous about eliminating this process because:

We need a fair and validated way to distribute compensation increases (donât we?)

We need a record of low performance when we let someone go

We need to capture performance data in an employeeâs profile for future promotion and other talent reviews, development plans, and career migration

We need a way to make sure managers are doing their jobs well.

Well, Iâve probably discussed these issues with 100+ companies over the last five years and our research shows more and more that companies are ready to let this process go.

Organization structures have changed and companies need to be more agile. We have a shortage of key talent and the keys to success now focus on regular alignment, coaching, creating passion and engagement, and continuous employee development.

The new keys to success:

Develop a âfeedback-richâ culture and set of tools (often online, sometimes formal, often informal) that encourages all employees to give each other feedback. Tools from companies like Achievers, Globoforce, and most HR software vendors now enable and make this easy.

Separate the discussions about performance from discussions about potential and future career plans. Yes we need to evaluate people when raise time comes, but that can be a totally different conversation from.

Talk about performance regularly and let employees create their own goals on a regular basis. Force managers to provide ongoing feedback and teach them how to have honest conversations.

Assume that employees already know something about their own performance, and force them to self-assess. People tend to have a good idea of their own strengths and weaknesses â give them an open and positive opportunity to share it. That starts the dialogue about expectations and the match between their self-assessment and that of the organization.

Enable managers to assess performance regularly. Software teams now use Agile tools which evaluate code on a weekly basis. Managers should be giving people feedback regularly. If they learn to do this on a regular basis it will get easier and employees will learn to appreciate it.

Focus managers on hiring the best, so they build a team which strives for 100% high performers. This is never possible of course, but rather than assuming that 20% of your employees will perform poorly, spend more time on assessment, culture, and fit to make sure very few low performers make it into the organization in the first place.

Remember that everyone wants to succeed. If they arenât performing well itâs not necessarily their fault â the organization should take responsibility for helping them find a better fit if possible.

Set and reset goals frequently. Companies that set performance goals quarterly generate 31% greater returns from their performance process than those who do it annually, and those who do it monthly get even better results. This means employees get feedback on a continuous basis (most sales organizations work this way). Read our research for more details.

Beware of pay for performance plans. While many companies (particularly investment banking, sales) have large pay for performance plans, research showsthat these can create perverse behavior. People focus on their own goals at the expense of the organization. In sales related roles this process works well â when you move to customer service, engineering, and other âbuilderâ roles they can create problems.

Give your leaders a cultural framework and set of values to work from. Companies like Juniper and Deckers (Uggs shoes) focus very heavily on corporate values, forcing managers to hire and manage to these values. This makes selection easier and enables us to evaluate and coach people against higher level frameworks.

Invest in leadership development. Being a manager is a tough job. Managers at all levels struggle with selection, hiring, training, coaching, and evaluation. Give them time and tools to learn, a framework for feedback, and a continuous development process so they learn how to become better.

Reward talent âproductionâ not talent âhoarding.â If you pay your managers to âproduce output,â they will focus heavily on talent performance and evaluation. This may be a good thing, but ultimately companies thrive by building skills â so in addition to encouraging managers to produce good work, incent them to produce good talent as well.

Iâve had the opportunity to study HR and performance management for almost 15 years now, and this topic has been controversial for a long time. In todayâs globalized,talent constrained workplace itâs clear to me that the traditional appraisal has to change.

Businesses thrive on agility, speed, passion, and alignment. The process of driving and measuring performance has to do the same.

HLTH 240b Lecture Notes - Lecture 10: Analysis Of Variance, Random Assignment, Measuring Instrument

Document Summary

Get access

Related Documents

HLTH 240b Lecture Notes - Lecture 6: Nonparametric Statistics, Parametric Statistics, Statistical Power

HLTH 240b Lecture Notes - Lecture 5: Nonparametric Statistics, Statistical Inference, Descriptive Statistics

PS395 Chapter 11: Independent Multi-Factorial ANOVA

Related Questions