2014 HYS Regional Workshop workshop... · Workshop purpose and objectives Help people understand...
Transcript of 2014 HYS Regional Workshop workshop... · Workshop purpose and objectives Help people understand...
2014 HYS Regional Workshop
Using your HYS Results April/May 2015
Welcome!
• Introduce presenters • General housekeeping • Powerpoint and workbook • Find out who is here
Workshop purpose and objectives
Help people understand and use their HYS results: • Overview of 2014 Results • HYS background & administration • Statistical issues • Accessing your data • Results available on AskHYS • Talking about your results
Overview of 2014 Results
The rates of alcohol use among 8th and 10th grades has dropped by half. Nearly 11,000 fewer students are using alcohol compared to 2010.
Q. Had alcohol during the past 30 days?
0%
10%
20%
30%
40%
50%
60%
1998 2000 2002 2004 2006 2008 2010 2012 2014
6th Grade 8th Grade 10th Grade 12th Grade
Source: WSSAHB 1998 (spring), 2000 (fall), HYS 2002, 2004, 2006, 2008, 2010, 2012, 2014 (fall)
-24%
Alcohol Use The prevalence of alcohol use has declined significantly in all grades.
Marijuana Use Marijuana use did not change significantly from 2012 to 2014
0%
10%
20%
30%
40%
50%
1998 2000 2002 2004 2006 2008 2010 2012 2014
6th Grade 8th Grade 10th Grade 12th Grade
Year 10th Grade Marijuana
Use
1998 26.6%
2000 21.9%
2002 18.3%
2004 17.1%
2006 18.3%
2008 19.1%
2010 20.0%
2012 19.3%
2014 18.1%
Q. Used marijuana/hashish during the past 30 days?
Source: WSSAHB 1998 (spring), 2000 (fall), HYS 2002, 2004, 2006, 2008, 2010, 2012, 2014 (fall)
Perception of Risk from Marijuana Use Increasingly more students think using marijuana regularly is not risky
21%
35%
46%
0%
10%
20%
30%
40%
50%
2002 2004 2006 2008 2010 2012 2014
8th Grade 10th Grade 12th Grade
Q. Using marijuana regularly has no risk/only slight risk
21%
18%
27%
16%
9%
35%
21%
10%
34%
18%
8%
Sad or hopeless? Serious thoughts about suicide?
Attempt suicide?
8th Grade 10th Grade 12th grade
Mental Health Status More than one in four students had depressive feelings in the past year.
Over 100,000 youth (12-17 year olds) seriously considered suicide in the past year, which is about one in every six students.
NOTES: Depressive feeling: felt so sad or hopeless for two weeks in a row that they stopped doing usual activities.
Percent of students answering “Yes”
29%
16%
26%
14% 15%
7%
9%
2%
Had depressive feelings in the past year No depressive feelings in the past year
Substance Use and Mental Health Status Students who had poor mental health are also more likely to report substance use.
NOTES: Tobacco products include cigarettes and chewing tobacco.
Use Alcohol
Use Marijuana
Use Tobacco*
Use Pain Killers
Use Alcohol
Use Marijuana
Use Tobacco* Use Pain
Killers
Substance Use and Poor Academic Performance
37%
47% 52%
47%
20%
Use Alcohol Use Marijuana Use Tobcco Use Pain Killers to Get High
No Substance Use*
Low Grades
Low Grades
Low Grades
NOTES: * Did not use alcohol, marijuana, tobacco products (cigarettes or chewing tobacco), or pain killers in the past 30 days.
Low Grades
Washington State 10th graders who reported substance use are twice as likely to have low grades in school than those who do not use substances
Q.
Low Grades
Putting them all together, what were your grades like last year? Percent of students who report receiving “C”, “D”, or “F” mostly.
Use Alcohol
School Risk and Protective Factors Scale Scores “Protected”, “At Risk”
• School Opportunities for Prosocial Involvement
• Decisions about class activities, talk to
teacher one-on-one, work on special projects, get involved in extracurricular activities, class discussion involvement…
13%
25% 24%
6%
17% 15%
0%
5%
10%
15%
20%
25%
30%
Current cigarette …
Current alcohol …
Current marijuana …
Not Protected Protected
Protected by “School Opportunities” ? Washington 10th graders who are protected by “School Opportunities” have lower substance use rates.
4%
16%
11%
3%
13%
25% 27%
7%
0%
5%
10%
15%
20%
25%
30%
Current cigarette
smoking*
Current alcohol
drinking*
Current marijuana
use*
Current other
drug use*.
Not at Risk At Risk
Washington 10th graders with the risk factor “Academic Failure” have higher rates of substance use than students without that risk factor.
At Risk for “Academic Failure”?
New Questions in 2014
Social/emotional learning
Marijuana questions
Texting and driving
Sexual orientation
Anxiety/worrying
Gangs at school
Stopping/reporting bulling
HYS Background & Administration
Past youth surveys
U Student Alcohol and Drug Use Survey W Washington State Adolescent Health Behaviors Y Youth Risk Behavior Survey H Healthy Youth Survey
U U W W W Y W H H H H H H
88 90 92 95 98 99 00 02 04 06 08 10 12 14
H
Statewide support for HYS
+ Schools, ESDs, Local Health & Community Partners!
2014 HYS participation
Over 223,000 students…
In all 39 counties…
• In 215 school districts…
• In 989 schools took HYS
Small School District Pilot
Could survey extra grades 7, 9 and 11
186 districts were eligible as small (150 or less in a grade)
All schools in the district could participate
Could get combined Middle (6/7/8) & High School (9/10/11/12) results
Goal = increase the number of respondents, smaller confidence intervals, more stable results
Pilot participation
81 districts and 141 schools in those districts participated
101 schools received combined middle and/or high school results
Interleaved survey forms for Survey
forms for 8th, 10th and 12th graders
Also for 9th and 11th
graders in the Pilot
Optional questions A-
enhanced and B-
enhanced
Survey form for 6th
graders
Also for 7th graders in the Pilot
Survey form content
Form A ½ students
Core All
students
Form B ½ students
Form C All 6th graders
Using HYS Data
• How have you used HYS data in your school or community?
• What issues are you seeing in your school or community?
• What would you like to look at in your data or have you looked at to understand these issues?
Why is HYS important?
“Why do we do this survey?
Because students are willing to share what’s going on in their lives.
This isn’t filtered through parents or educators; it’s coming directly from students. And what they tell us is very important.
It gives us insights into what they’re thinking and what concerns them. From that, we are able to provide relevant programs and
services that can help them be more successful.”
- Randy Dorn, Superintendent of Public Instruction
HYS Press Release March 2015
Statistical Issues: Fundamentals & Understanding Your Results
Common questions you may get about HYS
1. Are youth telling the truth? Can we trust our HYS results?
2. Why take student instruction time to do a survey?
• What are common questions/concerns you get about the HYS???
What is reliability?
The extent to which a measure, procedure or instrument yields the same result on repeated trials. A survey item is reliable if it consistently produces the same results under the same circumstances.
Administration Procedures:
• Standardized administration procedures
• Importance of survey
• No names
• Only student sees their answers
• Students resource list
What is validity?
The degree to which the results are likely to be true, believable, and free of bias and can be generalized to a larger population. A survey item is valid if it accurately measures the concept it is intended to measure.
HYS uses questions from established surveys.
Data Cleaning Procedures, that remove respondents with:
• Inconsistent answers
• High level of substance use (all substances, every day)
• Dishonest
• Wrong grade
Generalizability
• Can we “generalize” from the survey to the population?
• Challenges to generalizability:
o Response rates (why 70%?)
o Are different for schools, districts, counties
o Need to consider how these challenges affect interpretation of the results
Do I need to generalize?
• Yes, if you want to apply the results to a larger population. o 8th graders in our district said….
• Yes, if you want to compare to others or results over time
• No, if you want to just describe the students surveyed, in that moment, without confidence intervals: o Students at our school who took the survey said….
What are Confidence Intervals?
• The confidence interval is the ± number after each item estimate
• A confidence interval accounts for the fact that the reported value is probably a little different than the true value for all of the students
• A 95% confidence interval, for example, means that we are 95% confident that the true value is within the ± range
• Confidence intervals are important when you generalize results to a larger population.
What do CIs look like?
Your survey results say 18.0% (± 2.0%) of Grade 10 students used marijuana.
18.0 – 2.0 = 16.0
18.0 + 2.0 = 20.0
Interpret as between 16.0% and 20.0% used marijuana and looks like:
0 5 10 15 20 25 30
CurrentMarijuana Use
Your survey results say 18.0% (± 5.0%) of Grade 10 students used marijuana.
18.0 – 5.0 = 11.0
18.0 + 5.0 = 23.0
Interpret as between 11.0% and 23.0% used marijuana and looks like:
0 5 10 15 20 25 30
CurrentMarijuana Use
What do CIs look like?
Why are CIs different sizes?
The size of a confidence interval is affected by: • Number of students.
o In general, the more students surveyed, the smaller the confidence interval.
• Inherent variability.
o If most students answer a survey question in the same way, then there is less variability. The more variable the answers, the wider the CIs.
Why do we need CIs if data are valid?
• CIs account for variability, NOT the validity of the data.
• CIs allow for the comparison of results to others and
to ourselves over time.
• To protect you---say, from exaggerated claims.
Determining statistical significance
• The probability that differences in results are not due to chance alone.
• When using 95% confidence intervals, a difference between two groups is considered statistically significant if chance could explain it only 5% of the time or less.
Comparing CIs: significant
Local students 25% ±5% (so range is 20% to 30%)
State students 36% ±3% (so range is 33% to 39%)
CI’s don’t overlap, so difference is significant
0 5 10 15 20 25 30 35 40
Local State
20% 30%
33% 39%
Comparing CIs: non-significant
Local students 25% ±5% (so range is 20% to 30%)
State students 28% ±3% (so range is 25% to 31%)
Local CI range overlaps and overlaps the point estimate, so the difference is not significant
0 5 10 15 20 25 30 35 40
Local State
20% 30%
25% 31%
Point estimate, 28%
CI overlap
0 5 10 15 20 25 30 35 40
Local State
Comparing CIs: inconclusive
Local students 25% ±5 (so range is 20% to 30%)
State students 32% ±3 (so range is 29% to 35%)
CIs overlap, but don’t include a point estimate, so you can’t tell if the difference is significant
20% 30%
29% 35%
Point estimate, 32%
CI overlap
Practice using confidence intervals
If you don’t have your own report, use these results
26.9% to 31.9% 35.4% to 43.8%
Tool for testing significance
If you need to know for sure, there is a “Tool” to test for significance at: www.AskHYS.net/Training The spreadsheet tests the difference between two point estimates and their 95% CI to compute a p-value. • If your p-value is less than 0.05, then your difference
is significant. • Only use this test if you have at least 30 students. • Don’t use this test if you have 0% or 100%
Example – Comparing to the state
Since we couldn’t tell if there was a difference by using our CIs – let’s try the “Tool”: • Your students: 25.0% (±5.0) • Students statewide: 36.0% (±3.0)
P-value is less than 0.05, so there is a significant difference.
25.0
36.0
5.0
3.0
Your students
Statewide
0.0002177
Using results with large confidence intervals
If you are a small school, district or county your CIs will be large... • The estimate may be good, but only of the
students who took the survey this year.
CIs don’t take participation rates into account: • Low participation – there may be additional
bias • High participation – you can feel more
confident in your point estimates.
Questions?
Getting Access to AskHYS
Levels of access
County, ESD and State results are available to everyone.
School district and school building HYS results are available only with the permission of the superintendents’ offices.
Log on for school district and building access
Use your EDS login and password
Contact your District Data Security Manager for appropriate permissions to your building or district data
Successful login
Frequency Reports
Opening a frequency report www.AskHYS.net/Reports
Frequency report title and sections
• Introduction & Overview • Highlights • Selected Results by Gender • Understanding Your Report • Frequency Results • Risk & Protective Factor
Scales • Risk & Protective Factor
individual questions • Item Index • Core Questions
Intro - Survey participation
Submitted a survey
Submitted a “usable” survey
Students enrolled
Enrolled ÷ usable surveys
Pilot Combined High School Report participation rates
Report highlights
Selected results by gender
Smallest “n” example
A minimum n = 40 is needed if:
50% of the local respondents are
female and 50% are male
50% X 40= 20 females
50% X 40= 20 males
50.0% of females smoke
cigarettes and 50.0% of males
smoke cigarettes
50.0% X 20 = 10 females
50.0% X 20 = 10 males
% in each cell
Female Male
None 50% 50%
1 or more 50% 50%
n in each cell
Female Male
None 10 10
1 or more 10 10
Understanding Your Report
“N” is shorthand for “number” --- the number of students who took this survey or who answered a survey question. There are multiple reasons why different questions have different “N’s”: • Question is only on Form A or B • Question is optional • Question is near the end of the survey • Question is difficult to answer, or too
personal
Samples of n’s from different survey questions for a district with 400 10th graders:
N for a core question (all/most students)
N for question on Form A only (half the students)
N for optional question only on Form B (half the students, and some schools requested surveys without option questions)
N for near the end of the survey
17. Have you ever, even once in your life: Used marijuana? (n=398)
a. Yes 49.0% (±3.0)
b. No 51.0% (±3.0)
18. Have you ever, even once in your life: Used inhalants? (n=200)
a. Yes 11.0% (±4.0)
b. No 89.0% (±4.0)
100. How old were you when you had sexual intercourse
for the first time?† (n=100)
a. I have never had sexual intercourse 70.0% (±8.0)
b. 11 yhears old or younger 2.0% (±2.0)
c. 12 years old 2.0% (±2.0)
d. 13 years old 5.0% (±2.0)
e. 14 years old 8.0% (±6.0)
f. 15 years old 9.0% (±5.0)
g. 16 years old 3.0% (±3.0)
h. 17 years old or older 1.0% (±1.0)
150. Does your school have a counselor? (n=375)
a. Yes 90.0% (±3.0)
b. No 5.0% (±2.0)
c. Not sure 5.0% (±2.0)
Frequency results
18.0% (±2.0) 26.7% (±2.2)
82.0% (±2.0) 73.3% (±2.2)
10.0% (±1.0) 8.5% (±1.0)
6.0% (±1.0) 5.3% (±0.6)
1.0% (±0.5) 2.6% (±0.5)
1.0% (±0.5) 10.3% (±1.3)
(n=400) (n=9,000)
a. None
b. 1-2 days
c. 3-5 days
d. 6-9 days
e. 10 or more days
Any use in past 30 days
29. Use marijuana or hashish (weed, hash, pot?)
Between 16% and 20%:
• 18% - 2% = 16%, low
• 18% + 2% = 20%, high
During the past 30 days, on how many days did you:
Other frequency results
Pilot Multi-grade Middle School results
10.0% (±4.0) 8.0% (±1.0)
90.0% (±4.0) 92.0% (±1.0)
5.0% (±3.0) 3.5% (±1.0)
3.0% (±2.0) 2.5% (±0.5)
1.0% (±0.5) 1.0% (±0.2)
1.0% (±0.5) 1.0% (±0.2) 1
(n=100) (n=300)
a. None
b. 1-2 days
c. 3-5 days
d. 6-9 days
e. 10 or more days
Any use in past 30 days
29. Use marijuana or hashish (grass, hash, pot?)
2.0% (±2.0) 4.0% (±3.0)
98.0%(±2.0) 96.0% (±4.0)
2.0% (±2.0) 2.0% (±1.0)
0.0% (±0.0) 2.0% (±1.0)
0.0% (±0.0) 0.0% (±0.0)
0.0% (±00) 0.0% (±0.0) 1
(n=100) (n=100)
Pilot combined Middle School results
Grade 6 Grade 7 Grade 8 MS Combined
% (±CI) % (±CI) % (±CI) % (±CI)
8.0% (±1.0) 5.2% (±0.9)
92.0% (±1.0) 94.8% (±0.9)
3.5% (±1.0) 2.3% (±0.4)
2.5% (±0.5) 0.9% (±0.2)
1.0% (±0.2) 0.6% (±0.2)
1.0% (±0.2) 1.4% (±0.3) 1
a. None
b. 1-2 days
c. 3-5 days
d. 6-9 days
e. 10 or more days
Any use in past 30 days
29. Use marijuana or hashish (grass, hash, pot?)
Your Students Small Statewide
% (±CI) % (±CI)
Why are there asterisks?
• Results suppressed at school level to protect anonymity
• School chose not to ask optional question • Results by gender require a minimum number of respondents
Risk and protective factors
Each “factor” is a “scale” measured with two or more questions, so that the factor includes multiple dimensions of the risk or protection.
Risk factors that predict youth substance use, violence and delinquent behaviors. Protective factors that can protect youth from the effects of those risks.
Prevention strategies can be designed to lower risk and increase protection.
E.g., Academic Failure What grades do you usually get How far do you think you will get in school
Risk & Protective Factors Over Time
Report extras
Questions by topic
Core questions
PowerPoint Slides
Opening PowerPoint slides
Participation rates provided
Example of a typical slide
Example of a Pilot typical HS slide
Fact Sheets
Opening a fact sheet
Opening Fact Sheets
2014 Fact Sheet chart
Trend chart and table
State comparison chart
*
Pilot state comparison MS chart
*
Statewide relationship charts
All grade Fact Sheet charts
All grade Fact Sheet statewide relationship charts
Comparisons over time
A few things to think about first:
• Did the survey question change?
• Were challenges to generalizability similar? o Response rates o Groups of students missing
• Do you expect a change?
Trends
• Fact sheets do provide data back to 2004 for some questions – but do not provide a test for overall change in trend (only year to year comparisons)
• To measure trends you need statistical software o Free software from the National Cancer
Institute – uses your point estimates and confidence intervals:
http://surveillance.cancer.gov/joinpoint
Questions?
QxQ Online Data Query System
Q x Q = Question by Question
Running frequencies and crosstabs • Run frequencies: responses to individual
questions. Just like the results in your Frequency Reports.
• Run crosstabs: looking at the relationship between two questions - crossing one question by the second question.
Running crosstabs – HYS requirements
Cell size • State level minimum 5 per cell. • Sub-state level minimum 10 per cell. **NOTE** • The squares that have results in them are the
“cells”. • The number of respondents in each cell is
called the “n”, or also the “cell size”.
Crosstabs – cell size error message
“At least one cell in the results table contained a count of less than 10.” Output is suppressed.”
Cell size limitations can be frustrating! If your results are suppressed because of cell
size: • Select the “collapsed” answer option instead of
the “surveyed” option. • Re-run your crosstab at a higher geography – like
county rather than district.
Opening the Q x Q www.AskHYS.net/Analyzer
Opening the Q x Q www.AskHYS.net/Analyzer
Running a frequency - question selection
Frequency - response selection
Frequency - drag and drop
Frequency output
Running a crosstab – question selection
Alcohol and marijuana example Row Variable = Current Alcohol, Column Variable = Current Marijuana
Year=2014, Grade=8 Among 8th graders who didn’t drank: • 3.6% used marijuana (cell 2)
Among 8th graders who drank: • 49.4% used marijuana (cell 4)
Current Alcohol Drinking and Current Marijuana Use
96.4% ±0.7%
9,1850
3.6% ±0.7%
345
50.6% ±3.9%
421
49.4% ±3.9%
411
100.0%
9,530
100.0%
832
2014
Correct way: Race and bullying
Row Variable = Race, Column Variable = Bullying Year=2014, Grade=8
Among White non-Hispanics: • 29.1% were bullied
Among Hispanics: • 26.3% were bullied
Ra
ce
Bullied
Wrong way: Race and bullying
Row Variable = Bullying, Column Variable = Race, Year=2014, Grade=8
Among White non-Hispanics: • 51.3% were bullied
Among Hispanics: • 16.5% were bullied
Bu
llie
d
Race White non-
HispanicHispanic
American
Indian or
Alaska
Native non-
Hispanic
Asian or
Asian
American
non-Hispanic
Black or
African
American
non-Hispanic
Native
Hawaiian or
other Pacific
Islander non-
Hispanic
Other non-
Hispanic
Multiracial
non-HispanicTotal
48.7% 18.1% 3.0% 11.0% 4.1% 1.7% 7.6% 5.8% 100.0%
± 5.8% ± 5.3% ± 0.7% ± 3.5% ± 1.2% ± 0.4% ± 0.8% ± 0.8%
3,672 1,363 226 831 312 130 572 434 7,540
51.3% 16.5% 4.0% 7.0% 7.0% 1.5% 8.8% 7.0% 100.0%
± 5.0% ± 5.1% ± 0.9% ± 2.2% ± 1.0% ± 0.5% ± 1.2% ± 1.4%
1,509 486 117 205 114 45 259 205 2,940
no
days
any
days
Check you Q x Q results
You can compare your Q x Q frequency results with your Frequency Reports (www.AskHYS.net/Reports) • Page 4 in your reports includes crosstabs by gender Ask someone to review your results **NOTE** There may be slight differences because of the way different statistical programs round numbers.
Questions?
Crafting a Communication Message Around HYS
Talking about survey results
Keep it simple Round to a whole number
Survey results are estimates
Say “about”
Acknowledge low participation – “among those who took the survey”
Carefully include your confidence interval, if at all
If 75.3% (±5), between 70 and 80 percent
Plus or minus 5 percent
Different ways to talk with numbers
Painful: 75.3% (±5.1) of 8th graders feel safe at school.
Less painful: About 75 percent
About 3/4ths
About 3 out of 4
Turn it around, if that is the point you need to make: About 25% of 8th graders don’t feel safe
1 in 4
If you have 200 8th graders, turn it into the number of students, about 50 of our 8th graders don’t feel safe
Use visuals
There are charts in your PowerPoint slides and fact sheets
Create other types of charts or visuals:
7 ½ out of 10 students…
Where do you start?
Know your facts about HYS
How, when and to whom it was administered
Details about the questions
What information do you have?
HYS reports, slides, fact sheets
What do your 2014 HYS results say?
What HYS questions address your topic?
How do your results compare to the state results or results from previous years?
Where do you start?
Do the numbers make sense?
What else is needed to tell the whole story? Additional data sources
Information from informed people
Define your audience
Ask yourself: “Are there specific opinions I need to influence?”
Administration
Parents
Reporters
Grant application reviewers
What’s your Single Overriding Communication Objective (SOCO)?
This is the “so what” of your message. Ask yourself:
“If I could only make one point, what would it be?”
Tells your audience what you want them to learn or do
Practice: Using your results
Pick a scenario below and use HYS on page 60 of your Workbook to develop a SOCO or a key message.
The school board has invited your team to talk with them and make your case about the need for more early prevention programs among youth.
The son of a prominent city council member was arrested for driving under the influence of marijuana. He attends the local high school. A radio station calls and asks you to provide a sound bite about marijuana use among youth in your community.
A local agency is offering funds for programs that support youth. Support your grant application for a mentoring program.
Share Your Message
Share your results with the group
Example of SOCOs
“The school board needs to take action to prioritize marijuana prevention programs for our middle school children”
“Every parent in our community should talk to their kids about marijuana and driving”
“Funding this program will reduce the excess danger that children in our community face from marijuana”
What ideas do you have for using HYS data after leaving here today?
Reflection
Questions?
Thank you!
From the Healthy Youth Survey Planning Committee
• DASA-DBHR: Martha Perla, Linda Becker, Steve Smothers,
Grace Hong
• OSPI: Krissy Johnson, Dixie Grunenfelder
• DOH: Kevin Beck, Anar Shah, Lillian Bensley, Cathy Wasserman, Riley Peters
• Liquor Control Board: Mary Segawa
• Looking Glass: Joe Kabel and Susan Richardson
Questions?
About this training:
Krissy Johnson: [email protected]
Susan Richardson: [email protected]
Other HYS questions:
www.AskHYS.net