add something on publication bias in meta-analyses ...

[lizzieinvancouver]

We should say that ...

• publication bias can lead to both positive results being published more often and larger effect sizes being published (cite Kicinski_2018 or such).
• there are methods that can try to diagnose these, such as funnel plots, which have pros and cons (cite some of the papers I uploaded)
• we did not use these methods given that: (1) our design is slightly more complicated in that our effect size variation can also depend on the other treatments (i.e., variation in effect size of forcing depends on chill and photo levels); (2) we had limited sample sizes (give RANGE) and most often sample size was not reported.
• our partial pooling will pool extreme effects based on sample size in our database.

[lizzieinvancouver]

Should also add something about we do know the physiological basis for these cues generally so most studies have positive results for some of them (e.g., every study I know of finds a forcing and chilling effect ...).

[lizzieinvancouver]

@AileneKane I added some text to to supp section, 'The Observed Spring Phenology Responses in Experimental Environments (OSPREE) database.'

Still need to:

1. Add sample size range and % of studies with sample size.
2. Update the PRISMA checklist and where we refer to the numbers of it we did.

[lizzieinvancouver]

Done.