Baby Books Analysis Plans

Interocular trauma tests This is basically to assess what we have, what’s usable, what needs to be double checked, and what possibly needs transformations.Testing.

  1. Completeness/missingness by item, instrument, and participant
  2. Scale Summaries – Factor Structures (confirmatory factor analysis, using subscales provided by publishers) (see Baby Books\Measures\Measures_Simplified_With_Links_To_Codebooks.xls for details)
  3. Scale psychometrics (see Baby Books\Measures\Measures_Simplified_With_Links_To_Codebooks.xls for details)
  4. Means, medians, standard deviation, skew, kurtosis, outliers, reliability (where applicable)
Variable calculation/ key variables These are variables we need to address the specific aims that are yet to be calculated

Implementation
  1. Self-report reading of books
  2. Other book give aways
  3. Other educational sources
  4. Parenting classes
  5. Nurse home visitation
Medical Adherence, from Incurred Cost Interview
  1. immunization schedule
  2. well-child visits
Feeding Recommendations Adherence, from Infant Nutrition Interview
  1. Within-wave adherence
    1. food introductions (timing w.r.t. baby's age)-- calculated at CEPI
    2. food discontinuation (timing w.r.t. baby's age)-- calculated at CEPI
    3. various food amounts per week-- in process at CEPI
  2. Across-wave adherence
Threshold scores
  1. CES-D (depression): AWCESS
  2. W1 score of 8 or higher (baseline)
  3. Increase of 5 points or more between any two consecutive waves
  4. Two consecutive scores of 8 or higher
  5. At-risk on BINS (at any wave)
  6. At risk in any category: OWBSTARC, OWBSTARR, OWBSTARE, OWBSTARF, OWBSTARG
  7. Total number of categories in which baby is at risk: OWBSTRsk
  8. Atypical range on PLS (at any wave): OWPL03bF, OWPL06bF, OWPL09bF, each in conjunction with OWPLAcUp
Other scores (no particular threshold)
  1. NEO-FFI domain scores-- deciles from charts in manual: AWNEOD_N, AWNEOD_E, AWNEOD_O, AWNEOD_A, AWNEOD_C
  2. AAPI-- Sten scores from chart in manual: AWAPI_A, AWAPI_B, AWAPI_C, AWAPI_D, AWAPI_E
Knowledge about Development ("Opinions About Babies")
  1. Create "taught" and "untaught" subscales for each wave
  2. (Possibly) Match pre and post items
Group Equivalence Across 3 Groups at Wave 1 This is test whether our groups are equivalent at baseline. Need to compare VUBook, Commercial, NoBook on the following characteristics (assuming random assignment worked well)

  1. Demographics
    1. Age: AWMomAge
    2. race/ethnicity: AWBQ02E1-E6, AWBQ02ES, AWBQ03
    3. education: AWBQ04
    4. family income in last year: AWBQ17
    5. personal income in last week: AWBQ09, AWBQ13, AWBQ14, AWBQ16a, AWBQ16b, AWBQ16c
    6. employment: AWBQ09, AWBQ10, AWBQ11, AWBQ12, AWBQ16a
    7. public assistance: AWBQ18a-18g, AWBQ18gS
    8. marital status: AWBQ06
    9. Wonderlic total score: AWWPTot
    10. any others to be provided by SR
    11. Residence at W1? Miles form VU or scan of zip codes (would require going through participant files by hand)
  2. Baseline scores
    1. knowledge: AWO1BTC, AWO1BPC1, AWO1BPC2
    2. reading self-efficacy: AWRSPS
    3. reading beliefs: AWPRPT, AWPRPS1, AWPRPS2, AWPRPS3, AWPRPS4, AWPRPS5
    4. pregnancy experience: AWPESUF, AWPESUI, AWPESHF, AWPESHI, AWPESFR, AWPESIR
    5. depression: AWCESS
    6. breastfeeding intentions/beliefs: AWBI01, AWBITC, AWBIPC1, AWBIPC2
    7. any others to be provided by SR
Specific Aims to be addressed once all waves are merged

1 To test whether embedding anticipatory guidance in baby books is an effective method for increasing maternal knowledge of child development and parenting strategies. i.e., Are there differences between groups and across items (material covered or not) on the Opinions about Babies questions? – each wave and across study (i.e., Whether group predicts knowledge score)

  • Core covariates
    1. Mom’s Age at W1 ivw: A1MomAge
    2. Number of kinds of public assistance at W1 (A1BQ18T)
    3. Education at W1 (A1BQ04)
    4. Race/ethnicity (AWBQ02) and Hispanic (AWBQ03)
    5. Marital status at W1 (A1BQ06)

  • Specific covariates for this aim
    1. Aptitude (Wonderlic): A1WPTot
    2. Parenting classes at or before W1: A1BQ40
    3. Nurse visitation program at or before W1: A1BQ44
    4. Knowledge items percent correct untaught at W1: A1O1BPC2

  • Outcome Measures
    1. Opinions about Babies (knowledge scores)
2 To determine if increases in knowledge of child development and parenting strategies results in changes in parenting behavior. (Group -> knowledge -> behavior change)

  • Predictor: Knowledge by topic. One-time-only variables (all waves after W1 combined; "taught" items (both kinds) only):
    • ANOBDeT: OB Development subscale: tot # correct
    • ANOBDeA: OB Development subscale: tot # asked
    • ANOBDeP: OB Development subscale: pct correct, based on # asked
    • ANOBMaT: OB Matrnl Hlth subscale: tot # correct
    • ANOBMaA: OB Matrnl Hlth subscale: tot # asked
    • ANOBMaP: OB Matrnl Hlth subscale: pct correct, based on # asked
    • ANOBNuT: OB Nutrition subscale: tot # correct
    • ANOBNuA: OB Nutrition subscale: tot # asked
    • ANOBNuP: OB Nutrition subscale: pct correct, based on # asked
    • ANOBPaT: OB Parenting subscale: tot # correct
    • ANOBPaA: OB Parenting subscale: tot # asked
    • ANOBPaP: OB Parenting subscale: pct correct, based on # asked
    • ANOBSaT: OB Safety subscale: tot # correct
    • ANOBSaA: OB Safety subscale: tot # asked
    • ANOBSaP: OB Safety subscale: pct correct, based on # asked

  • Core covariates: Same as for Aim 1

  • Other variables related to this aim
    1. Infant health:
      • SerMedProb: Reason to believe baby has serious medical problm? (yes/maybe/no) (one-time var-- all waves combined) Note that of the 168 babies whose mothers had at least one post-W1 interview, there are only 3 yes's and 6 maybe's for this item-- the rest are no's.
    2. Infant temperament:
      • 9 subscales on ITQ (W5 only): AWITActv AWITRthm AWITAppr AWITAdap AWITIntn AWITMood AWITPrst AWITDist AWITThrh

  • Outcomes: Parenting Behavior (listed below by type)
    1. Safety Practices
      • Measures
        • Home Safety Assessment (W2, W3, W4, W5, W6, W7)
          • Risk in immediate environment:
            • AWHSRskT (total # risk categories; high= bad)
            • AWHSRskA (# of risk categories asked/monitored)
            • AWHSRskP (% risk, based on # categories asked/monitored; high = bad)
          • Active safety practices:
            • AWHSActT (total; high= good)
            • AWHSActA
            • AWHSActP (%; high = good)
            • Note that there is some overlap between these vars and the risk vars above.
        • Infant Nutrition Ivw (W2, W3, W4, W5, W6, W7)
          • AWINChki: M refrained from giving certain choke hazards (note that AWINPC (used below) encompasses this variable)
          • ANINHazT: Tot # waves in which M reported feeding certain choke hazards to baby (one-time-only var)
          • AWIN26f: No honey given if baby under 1 yr old (note that AWINPC (used below) encompasses this variable)
        • Incurred Cost Interview
          • AWICInjW: Did mom report that baby had one or more injuries in this wave? (y/n)
          • ANICInjT: Total # of waves in which mom reported that baby had an injury (one-time-only var)
        • Chart audit
          • ZNSum1: Total # of injuries found in baby's medical records (one-time-only var). You may also want to take ZNSum9 into consideration: Did we receive all the medical records we requested for this child? (y/n)
      • Additional specific Covariates for this parenting behavior
        • A1HSRskP: W1 value of AWHSRskP
        • A1HSActP: W1 value of AWHSActP
        • A1O1BSaT, A, P: W1 knowledge items- safety subscale- Total Correct, # asked, Percent correct based on # asked
    2. Breastfeeding Practices
      • Measures
        • Breastfeeding Practices Q.
          • ANBPEver: Did mom ever breastfeed baby? (y/n): (one-time-only var)
          • ANBPWks: # of weeks mom breastfed baby: (one-time-only var)
          • AWBP24Hr: Number of times per 24 hours period (for those not breastfeeding = 0) (W2, W3, W4, W5, W6, W7)
      • Additional specific Covariates for this parenting behavior
        • Breastfeeding intention at W1: A1BI01
    3. Reading Frequency
      • Measures
        • Home Literacy Questionnaire
          • AWHLQ03: how many times a week do you read to your child (all waves 2-7 & A-L)
      • Additional specific Covariates for this parenting behavior
        • A1PRPT: Prenatal Parental Beliefs about Reading Inventory summary score
        • A1HLQ03: W1: how many times a week do you read to your child?
    4. Nutrition
      • Measures
        • Infant Nutrition Ivw
          • Age-of-introduction/feeding variables: (W2, W3, W4, W5, W6, W7)
            • AWINRef1: M refrained from giving certain babyfood/solids before OK age
            • AWINRef2: M refrained from giving certain liquids before OK age (including whole-fat cow milk, which is one of the types of milk included in AWINRef3)
            • Note that the above two vars do NOT take into account food/drinks that baby was never supposed to have in the first place.
            • AWINRef3: M refrained from giving any cow milk before OK age
            • AWINMilk: M gave age-approp breastmilk/formula/milk
            • AWINFrut: M gave age-approp fruit
            • AWINVegt: M gave age-approp veg
            • AWINFrVg: M gave age-approp fruit AND veg (Note: this one is a summary of the previous two—it equals “yes” if and only if both of them equal “yes”)
            • All of the above variables are yes/no.
            • AWINTC: Tot # correct nutrition behaviors (not including 'dosage') based on age
            • AWINTA: Tot # applicable nutrition behav. (not including 'dosage') basd on age
            • AWINPC: Pct correct nutrition behaviors (not including 'dosage') based on age
            • Note that these three are NOT independent of the yes/no vars above
          • Volume-of-food variables
            • AWINEnFl: Enough fluids y/n (W2, W3, W4)
          • One-time-only adherence vars:
            • CerFirst—did the mom start out with baby cereal before moving to baby food? (calculated by looking at all waves for each mom)
      • Additional specific Covariates for this parenting behavior
        • none
    5. Health Maintenance
      • Measures
        • Incurred Cost Ivw
          • Adherence to immunization guidelines: (one-time-only var)
            • ANICImmT: SelfRpt: Total # imm, adjusted
            • ANICImmP: SelfRpt: Adj Tot # imm as pct of recommended
            • ANICImWT: SelfRpt: Total # imm in window, adjusted
            • ANICImWP: SelfRpt: Adj Tot # imm in win. as pct of recommended
        • Chart audit
          • Four similar vars for each mom
            • ZNVCImmT: VacCard: Total # imm, adjusted
            • ZNVCImmP: VacCard: Adj Tot # imm as pct of recommended
            • ZNVCImWT: VacCard: Total # imm in window, adjusted
            • ZNVCImWP: VacCard: Adj Tot # imm in win. as pct of recommended
            • Note that the above 4 variables are set to 'missing' if we have reason to believe that we do not have full immunization records for that ID
          • Adherence to well-child-visit guidelines:
            • ZNSum3: Baby had 2-month checkup (54 - 68 days)
            • ZNSum4: Baby had 4-month checkup (115-129 days)
            • ZNSum5: Baby had 6-month checkup (176-190 days)
            • ZNSum6: Baby had 9-month checkup (267-281 days)
            • ZNSum7: Baby had 12-month checkup(358-397 days
            • ZNSum8: Baby had 18-month checkup(517-580 days)
            • ZNSum10: Total number of well-child visits (regardless of timing)
            • ZnSum11: Number of 'in-window' Well Check-Ups, 2 mos thru 18 mos (max = 6)
            • Note that ZNSum10 and ZNSum11 are set to 'missing' if we know that we do not have complete medical records for that ID
      • Additional specific Covariates for this parenting behavior
        • none
    6. Parenting Environment and Interactions
      • Measures
        • HOME inventory (W2, W5, W6, W7)
          • Subscales: each is the sum of yes/no items
            • AWHOMSR = Responsivity = AWHOM01C-11C subtotal
            • AWHOMSA = Acceptance = AWHOM12C-19C subtotal
            • AWHOMSO = Organization = AWHOM20C-25C subtotal
            • AWHOMSL = Learning Materials = AWHOM26C-34C subtotal
            • AWHOMSI = Involvement = AWHOM35C-40C subtotal
            • AWHOMSV = Variety = AWHOM41C-45C subtotal
          • Total score: AWHOMT [0,45]
        • Reading interaction video (W2, W3, W4, W5, W6, W7)
          • (variables to be specified later)
        • Toy play interaction video (W2, W3, W4, W5, W6, W7)
          • (variables to be specified later)
        • Home Literacy Questionnaire
          • AWHLQ08 (TV use; lower = better) (W2, W3, W4, W5, W6, W7)
      • Additional specific Covariates for this parenting behavior
        • none
Blue text: Second priority

3 To assess whether benefits of reading coupled with increased knowledge result in higher quality mother-child interactions. Group and knowledge -> better interactions (expect VU group to be highest, then comp book group, then no book group)

  • Core covariates: Same as for Aim 1

  • Other variables related to this aim (SR would like to use as covariates, but they are not available at baseline)
    1. Self-reported reading: AWHLQ09, 10, 11 (values: number of times mom read) (all waves 2-7 & A-L, but only for Groups 1 & 2).
    2. Infant Health: See Aim 2
    3. Infant temperament: See Aim 2
  • Outcome Measures:
    • Toy play videos to be added later (that is, in future years; will not be included in the current analysis)
    • Reading Interaction videos (W2, W3, W4, W5 to be added, W6, W7)-- commercial books (OWRIVidType = "B") only
      • Quality of Interaction: S_owrin41VidB (integer values 0-4)
      • Maternal warmth:
        • Per minute: OWRIQ2MVidB
      • Maternal negativity (higher values = more negativity):
        • Per minute: OWRIQ3MVidB
      • Maternal responsivity:
        • Per minute: OWRIQ4MVidB
      • Duration of reading episode, in seconds: T_owrinTmVidB
      • Quality of voice: S_owrin30VidB
      • Voice Effort(Intonation/Melody): S_owrin6VidB
    • HOME Inventory (W2, 5, 6, 7)
      • Emotional and Verbal Responsivity of Mother: AWHOMSR
      • Involvement: AWHOMSI
      • Acceptance: AWHOMSA
Purple text: Fourth priority

  • Possible SEM:
    • Latent Construct = Maternal Responsivity Indicators = HOME responsivity subscore, Reading Videos responsivitiy score; Warmth = maternal warmth and HOME acceptance

Orange text: Third priority

4 To determine if increased knowledge and improved maternal-child interactions result in less maternal stress and increased maternal self-efficacy.
  • Core covariates: Same as for Aim 1

  • Other variables related to this aim (SR would like to use as covariates, but they are not available at baseline. From SR: "Basically, these are conceptually interesting but are endogenous to the study. We need to look at the relationship without these but I would like an analysis that includes these to better understand how these social aspects influenced outcomes perhaps mediationally.")
    1. Social Support:
      • Other adults in home:
        • AWBQ07S1 (yes/no)
        • AWBQ07S2 (0=no in-home support, 2=b's dad, 1=other)
      • Support in neighborhood
        • AWCISup: Support in nhood (0=none, 1= relative(s) or friend(s), 2= both)
        • AWCI07: participation in groups in neighborhood – y/n
        • AWCI07T: neighborhood participation score = sum of AWCI07A-F, M
        • Each mom should have a "first" CI (Community Conditions Ivw) done at W3 or asap thereafter, and then one or more other CI's if she moved after the first one. How do we want to handle this?
    2. Infant Health: See Aim 2
    3. Infant temperament: See Aim 2

  • Outcome Measures
    • Parenting Stress Index (W2, W4, W6, W7)
      • variables: total score (AWPSITS: higher score = more stress); could look at subscores if interesting i.e., AWPSIDR = PSI Defensive Responding score, AWPSIPD = PSI Parental Distress subscale, AWPSIDI = PSI Parent-Child Dysfunctional Interaction subscale, AWPSIDC = PSI Difficult Child subscale
      • Additional specific Covariates for this measure:
        • none
    • Maternal Self-Efficacy (W2, W3, W5, W6, W7)
      • variables: total score (AWMSES)
      • Additional specific Covariates for this measure:
        • none
    • Parental Satisfaction (W2, W3, W4, W5, W6, W7)
      • variables: total score (AWPSSS)
      • Additional specific Covariates for this measure:
        • none
    • CES-D (W2, W3, W4, W5, W6, W7)
      • variables
        • AWCESS: total score
        • AWCESDep: is AWCESS >=10? (y/n) (if yes, mom is considered depressed)
      • Additional specific Covariates for this measure:
        • A1CESS (= AWCESS from W1)
Green text: First priority once all the EP video coding is done

5 To assess if better parenting practices, higher quality mother-child interactions, and reduced maternal stress result in better child outcomes (listed below).
  • Predictors
    • Better parenting practices:
      • Safety Practices from Aim 2
      • Attitudes: AAPI (W2, W4, W6, W7)
        • AWAPISA, AAPI Inapprop. Expectations score [7-35]
        • AWAPISB, AAPI Empathy score [10-50]
        • AWAPISC, AAPI Corporal Punishment score [11-55]
        • AWAPISD, AAPI Role Reversal score [7-35]
        • AWAPISE, AAPI Power Independence score [5-25]
        • AWAPI_A, AAPI Inappr. Expect. norm-ref. Sten scor
        • AWAPI_B, AAPI Empathy norm-ref. Sten score
        • AWAPI_C, AAPI Corporl Punshmt norm-ref. Sten scor
        • AWAPI_D, AAPI Role Reversal norm-ref. Sten score
        • AWAPI_E, AAPI Power Indep. norm-ref. Sten score
      • Beliefs: PRBI (W4, W6, W7)
        • AWPRBT, PRB total score
        • AWPRBS1, PRB Teachg Efficacy subscore
        • AWPRBS2, PRB Positive Affect subscore
        • AWPRBS3, PRB Verbal Participation subscore
        • AWPRBS4, PRB Reading Instruction subscore
        • AWPRBS5, PRB Resources subscore
    • Higher quality mother-child interactions: Reading Interaction outcome measures from Aim 3
    • Reduced maternal stress: Parenting Stress Index outcome measures from Aim 4

  • Core covariates: Same as for Aim 1

  • Other variables related to this aim (SR would like to use as covariates, but they are not available at baseline. From SR: "I think I would like to see [the analysis] with and without these variables as I think infant health is relevant to quality of interactions although we definitely need to look at it without this at it is endogenous.")
    1. Infant Health: See Aim 2
    2. Infant temperament: See Aim 2
  • Outcome Measures
    • Injuries: see Aim 2, under "Safety Practices"
    • Preschool Language Scale (PLS) (W4, W6, W7)
      • variables: OWPL03b1, OWPL06b1, OWPL09b1 (standard scores)
      • Other variables related to this measure: (SR would like to use as covariates, but they are not available at baseline. From SR: "it's relevant to capture measurement differences due to language at home rather than true language ability as measured by the PLS")
        • OWPLEngM (Is Engl. the lang. spoken most to child?); we have this info for each wave the PLS was administered. Also, LS can add “Other than when we visit your home, is English ever spoken to the child?” if there are enough “no’s” in answer to the above question to make it worthwhile.
    • Bayley Infant Neurological Screener (BINS) (W3, W4, W5)
      • variables: OWBSTRC, ...RR, ...RE, ...RF, ...RG (the Risk Category Scores for the 5 developmental areas tested); if there is no variability in this, then we can look at raw scores again.
      • Additional specific Covariates for this measure:
        • none
    • Exploratory Play video (W6, W7) Note: as of 1 June 2010, Toy 4 is not yet clean, and Toy 5 is not yet coded.
      • "Per toy" summary variables (N is replaced by toy number 1-5 in variable names):
        • HighestFAllSegsToyN: Highest 'F' code for this toy
        • FreqHighestFToyN: # times the highest 'F' code was coded at all for this toy (max=120)
        • HighestPAllSegsToyN: Highest 'P' code for this toy
        • FreqHighestPToyN: # times the highest 'P' code was coded at all for this toy (max=120)
      • Across-toys summary variables Note: Stephanie may want to change the calculation of these based on the number of toys completed by each baby:
        • AllToys_HighestF: Highest 'F' code across all toys
        • AllToys_FreqHighestF: # times the highest 'F' code was coded at all across all toys (max=600)
        • AllToys_HighestP: Highest 'P' code across all toys
        • AllToys_FreqHighestP: # times the highest 'P' code was coded at all across all toys (max=600)
      • Additional specific Covariates for this measure:
        • none
6 Assess the cost-effectiveness of the baby books intervention. This is all AR’s plan :) (note, you may want to include delayed second child in this)

Future Data Needs
  1. Conversion of ProCoderDV files This will make Aims 2 -5 feasible
    1. Reading
    2. Exploratory play task
    3. Mother-child interaction
  2. Analysis of ProCoderDV files
    1. by infant assessor
  3. Medical chart audit
  4. hospitalization follow-ups
Quick Paper Ideas Here are some ideas for some quick papers on the data was have merged so far, esp. is Ben is nervous about publishing/tenure
  1. Concordance of breastfeeding intentions and practices: Facilitator and barriers to nursing infants (This could be W1 and W2 or could include up to W6)
  2. Cross sectional study (including total n) of how women’s demographic characteristics, personality, and intelligence contribute to their pregnancy experience, knowledge of child development, breastfeeding intentions, and parental beliefs about reading. (SEM would fit well – although mainly descriptive)
  3. The relationship between of parenting attitudes and beliefs and child temperaments in the incidence of preventable injuries and illness (ala Morrongiello, Klemencic, & Corbett, 2008)
  4. Focus on reported facilitator and barriers (BF practices)
  5. Medical professor support/undermining (e.g., formula at hospital, no discussion of nursing)
  6. Demographic contributors (e.g., return to work, low education, etc.
Topic revision: r75 - 30 Jul 2010, LaurieSamuels
 

This site is powered by FoswikiCopyright © 2013-2017 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Vanderbilt Biostatistics Wiki? Send feedback